Integra V Providence Second Complaint
Integra V Providence Second Complaint
Integra V Providence Second Complaint
1 TABLE OF CONTENTS
2 Page
3
4 I. INTRODUCTION ............................................................................................. 1
5 II. JURISDICTION AND VENUE........................................................................ 2
6 III. PARTIES ........................................................................................................... 3
7 IV. SUBSTANTIVE ALLEGATIONS ................................................................... 4
8 A. Overview of Medicare Reimbursement and Upcoding .......................... 4
9 B. In Consultation with JATA, Providence Pushed Its Coders, CDI
Specialists, and Doctors to Apply Unnecessary CCs and MCCs ........... 6
10
1. Providence, aided by JATA, trained doctors to upcode
11 MCCs ............................................................................................ 6
12 2. With the Assistance of JATA, Providence Pressured
Doctors to Upcode MCC Using Leading Queries ...................... 10
13
3. JATA Gave Providence Tips on How to Avoid Audits of
14 Their Upcoding ........................................................................... 13
15 4. Providence and JATA Incentivized Staff to Upcode MCCs ...... 14
16 5. JATA’s software promoted upcoding MCCs ............................. 15
17 6. JATA’s upcoding strategies extend beyond Providence ............ 16
18 C. Relator’s Methodology ......................................................................... 17
19 D. Defendants’ False Claims ..................................................................... 20
20 1. The False Claims Made by Providence in Consultation
with JATA................................................................................... 20
21
(a) Encephalopathy ................................................................ 23
22
(i) Specific Patterns of Fraud with
23 Encephalopathy ...................................................... 25
24 (ii) Specific False Claims with Encephalopathy .......... 33
25 (b) Respiratory Failure ........................................................... 38
26 (i) Specific Patterns of Fraud with Respiratory
Failure .................................................................... 41
27
(ii) Specific False Claims with Respiratory
28 Failure .................................................................... 44
394151.1 ii Case No. 2:17-cv-01694-PSG-SS
SECOND AMENDED COMPLAINT
Case 2:17-cv-01694-PSG-SS Document 38 Filed 08/10/18 Page 3 of 100 Page ID #:644
1 employees confirm that Providence and JATA worked seamlessly to create a culture
2 throughout Providence that promoted increasing Medicare billing without regard for
3 accuracy. Providence and JATA’s efforts ranged from pushing doctors to make
4 unwarranted diagnoses, to using leading queries to change doctors’ original
5 diagnoses. Even JATA’s proprietary software was designed to manipulate
6 diagnoses to maximize Medicare revenue.
7 4. In addition to identifying the Defendants’ false claims through its
8 proprietary analysis—and then confirming its findings through exhaustive
9 investigation—Relator also performed extensive econometric analysis designed to
10 eliminate conceivable innocent explanations. Thus, for instance, Relator’s analysis
11 rules out the possibility that the Defendants’ inflated Medicare billings arise from an
12 issue with the treating doctors or the type of patient that Providence treats.
13 Moreover, to be conservative, only the most extreme, statistically significant cases
14 of upcoding have been identified by the Relator as fraudulent.
15 5. In short, Relator has determined that Providence, with the
16 encouragement of JATA, has submitted more than $188.1 million in false claims for
17 Medicare reimbursement over the past seven years.
18 II. JURISDICTION AND VENUE
19 6. This Court has subject matter jurisdiction over this action pursuant to
20 31 U.S.C. § 3732(a) and 28 U.S.C. § 1331.
21 7. This Court has personal jurisdiction over the named Defendants
22 because, inter alia, the Defendants transacted business in this District; reside in this
23 District; engaged in wrongdoing in this District; and/or caused the submission of
24 false or fraudulent claims in this District. Further, 31 U.S.C. § 3732(a) provides for
25 nationwide service of process.
26 8. Venue is proper in this District under 31 U.S.C. § 3732(a) and 28
27 U.S.C. § 1391(b) and (c). During the relevant time period, a substantial portion of
28 the events complained of that gave rise to Plaintiff’s claims occurred in this District
394151.1 2 Case No. 2:17-cv-01694-PSG-SS
SECOND AMENDED COMPLAINT
Case 2:17-cv-01694-PSG-SS Document 38 Filed 08/10/18 Page 6 of 100 Page ID #:647
10
11
12
13
14
15
16
17
18
19
20 29. The doctor tip sheets created by JATA and disseminated to Providence
21 doctors also reveal how doctors were trained to document specific diagnoses
22 identified by Relator for excessive use—encephalopathy, acute respiratory failure
23 and severe malnutrition—in order to get an MCC. For example, while
24 encephalopathy was noted as an MCC, other related diagnoses were discouraged
25 because they were “not even a CC.”
26
27 5
Office of the Inspector General, U.S. Department of Health & Human Services,
28 OIG Work Plan 2017 at 7 (2017), available at https://goo.gl/BsJPyZ.
394151.1 8 Case No. 2:17-cv-01694-PSG-SS
SECOND AMENDED COMPLAINT
Case 2:17-cv-01694-PSG-SS Document 38 Filed 08/10/18 Page 12 of 100 Page ID #:653
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15 30. Similarly, acute respiratory failure was highlighted as “an MCC,” yet
16 respiratory distress was also de-emphasized since it yielded “little credit.” With
17 respect to severe malnutrition, the tip sheet for surgeons instructs doctors:
18 “Document severe malnutrition—it not only adds severity as an MCC, it will likely
19 prolong the post-op course thereby aligning the illness severity with length of stay.”
20
21
22
23
24
25
26
27
28
394151.1 9 Case No. 2:17-cv-01694-PSG-SS
SECOND AMENDED COMPLAINT
Case 2:17-cv-01694-PSG-SS Document 38 Filed 08/10/18 Page 13 of 100 Page ID #:654
1
2
3
4
5
6
7
8
9
10
11
12 31. JATA’s own staff directly trained Providence doctors to code MCCs.
13 In 2015, a JATA Regional Director created a video training Providence surgeons
14 and anesthesiologists on how to manage the transition from ICD-9 codes to ICD-10
15 codes. When covering the diagnoses related to cholelithiasis, which has 40 possible
16 ICD-10 codes, the video instructed Providence that only the code resulting in an
17 MCC is the “appropriate” diagnosis. Referring to this MCC, the Regional Director
18 instructs, “This is the way we describe conditions anyways, pretty much, and this is
19 the only one that is going to get you a major comorbidity, by the way. So as long as
20 we describe things in this way we are going to arrive at the most appropriate
21 diagnostic code over here out of the 40 different possibilities.”
22 2. With the Assistance of JATA, Providence Pressured Doctors
23 to Upcode MCC Using Leading Queries
24 32. JATA also influenced Providence’s coding process in order to boost the
25 system’s “capture” of CCs and MCCs. Hospitals are not allowed to apply a CC or
26 MCC unless it is sufficiently documented in the patient’s medical files. Providence
27 CDI specialists thus sent “queries” to doctors designed to push them to change their
28 initial assessments in ways that would justify a coding of a CC or MCC.
394151.1 10 Case No. 2:17-cv-01694-PSG-SS
SECOND AMENDED COMPLAINT
Case 2:17-cv-01694-PSG-SS Document 38 Filed 08/10/18 Page 14 of 100 Page ID #:655
1 about JATA’s views on leading queries: “[JATA] informed my CDI team that since
2 we are all RN's and part of the clinical team we do not have to be as concerned
3 about leading clarifications as a coder would…I am having a hard time accepting
4 this because all the literature I have researched has said to give multiple choice
5 options and never give just one diagnosis option as that would be considered leading
6 the provider.” In another CDI forum discussion on leading queries, a CDI notes that
7 “[JATA] allows their CDS [clinical documentation specialists] more leeway” by
8 providing a different standard for nurses than coders. However, CDI nurses are not
9 clinicians, but are instead focused on documenting care for the purposes of medical
10 coding. As another CDI specialist points out, “As a nurse speaking to a physician
11 [asking about a specific diagnosis] would be considered appropriate communication,
12 but as a CDI Specialist speaking to a physician, it is considered inappropriate to
13 ‘lead’ that way.” Indeed, in one of the U.S. Department of Justice’s enforcement
14 actions against a hospital for upcoding, the CDI was even a physician, but
15 nonetheless culpable of advising treating physicians using leading queries.
16 37. Second, JATA unscrupulously justifies leading queries by convincing
17 hospitals that it is doctors, not CDIs, who are entirely responsible for
18 documentation. In one instance discussed in the CDI member forum, a CDI was
19 evaluating JATA CDI software and noticed that the JATA software presented a new
20 diagnosis that generated leading queries. When she raised this issue to JATA,
21 JATA responded that “[doctors] are responsible for their response and we are not
22 responsible regardless of the query.”
23 38. JATA’s argument that only doctors, not CDIs, are responsible for
24 documentation is particularly troubling when considering how they trained CDI and
25 doctors. JATA-trained CDI exerted pressure on doctors to document for higher
26 severity to the point of acquiescence. For instance, a former coder at JATA facility
27 recalls doctors being tired of being constantly queried to the point that the doctors
28 would just ask what they should right down, as opposed to using their clinical
394151.1 12 Case No. 2:17-cv-01694-PSG-SS
SECOND AMENDED COMPLAINT
Case 2:17-cv-01694-PSG-SS Document 38 Filed 08/10/18 Page 16 of 100 Page ID #:657
1 judgement to document accurately; there were “doctors that will just say whatever
2 you ask him to say just to get you off his back.” Furthermore, JATA’s Regional
3 Director taught Providence doctors that the only way to avoid being queried by a
4 CDI would be to document an MCC. This pressure to code MCCs would
5 sometimes result in the creation of contradictory medical records. According to a
6 former Providence CDI staffer, it was not uncommon for doctors to initially
7 document delirium, but after being queried they hastily added encephalopathy to the
8 bottom of the chart.
9 39. In sum, JATA advised CDIs to issue leading queries to doctors who
10 were responsible for the medical record. Knowing doctors wanted to avoid being
11 queried, JATA advised doctors they could avoid queries by documenting an MCC.
12 This fraudulent scheme resulted in the excessive coding of MCCs identified by
13 Relator through its proprietary statistical analysis.
14 3. JATA Gave Providence Tips on How to Avoid Audits of
15 Their Upcoding
16 40. All the while promoting leading queries and the excessive coding of
17 MCCs, JATA also coached hospitals to avoid being audited. For example, in a
18 presentation to CDI personnel at an industry event, Mario Perez, Director of Clinical
19 Consulting at JATA, advised hospitals to avoid single CC or single MCC
20 diagnoses—one of the metrics tracked in the PEPPER report issued by CMS to
21 support hospitals’ compliance activities—since that was a target identified by
22 Medicare auditors as being at high risk of improper payment.
23
24
25
26
27
28
394151.1 13 Case No. 2:17-cv-01694-PSG-SS
SECOND AMENDED COMPLAINT
Case 2:17-cv-01694-PSG-SS Document 38 Filed 08/10/18 Page 17 of 100 Page ID #:658
1
2
3
4
5
6
7
8
9
10
11
12
13 4. Providence and JATA Incentivized Staff to Upcode MCCs
14 41. To increase the coding of MCCs, Providence gave CDI and doctors
15 financial incentives for successful queries, and it made clear to doctors and staff that
16 it closely tracked their responsiveness. For instance, a monthly newsletter at one
17 Providence hospital posted “Physician Agree Rates” and thanked medical and
18 surgical staff for responding to queries. It also provided CDI specialists with what
19 the hospital deemed as that month’s “Top 5 Queries,” as well as the “Most Effective
20 Query” measured by how much that query helped the hospital’s bottom line.
21 Frequently among the “Top 5 Queries” were the three MCCS identified by Relator
22 for excessive use—encephalopathy, respiratory failure and malnutrition. CDI staff
23 thus had the incentive to query doctors to document these diagnoses since CDI
24 performance is measured by the financial impact of their queries.
25
26
27
28
394151.1 14 Case No. 2:17-cv-01694-PSG-SS
SECOND AMENDED COMPLAINT
Case 2:17-cv-01694-PSG-SS Document 38 Filed 08/10/18 Page 18 of 100 Page ID #:659
1
2
3
4
5
6
7
8 42. Providence measured its CDI specialists by their CC and MCC capture
9 rates, thus giving them an incentive to upcode. Notably, JATA’s software tracked
10 the net effect of both positive and negative queries. If a coder or CDI specialist
11 issued a negative query—i.e., a query that reduced the severity of a DRG—JATA’s
12 software reduced the net effect of that coder’s queries. Thus, Defendants not only
13 incentivized upcoding, but disincentivized coders from sending negative queries.
14 This helped to create a one-way ratchet for Medicare billing—and helped JATA
15 deliver on its guaranteed increase of Providence’s CMI.
16 43. Relator’s investigation also revealed evidence that Providence tied
17 doctor salaries to the extent to which they responded to CDI queries. A former
18 Providence CDI specialist recalls that doctors employed by Providence would have
19 their performance and salary negatively impacted if they did not respond to queries.
20 5. JATA’s software promoted upcoding MCCs
21 44. Providence used JATA’s proprietary software to help design queries.
22 Using JATA’s software, the moment a CDI specialist entered the patient’s principal
23 diagnosis, the CDI specialist was told immediately if a CC or MCC would impact
24 the severity level of the DRG. The software then gave the user suggestions of
25 possible CCs and MCCs along with associated signs and symptoms.
26 45. JATA clinical documentation software has a reputation for generating
27 outsized additional revenue. In a CDI discussion forum, a CDI manager recounted
28 how her new VP of Finance used JATA in the past and wanted her department to
394151.1 15 Case No. 2:17-cv-01694-PSG-SS
SECOND AMENDED COMPLAINT
Case 2:17-cv-01694-PSG-SS Document 38 Filed 08/10/18 Page 19 of 100 Page ID #:660
1 switch to JATA. Despite having a “mature CDI program” that had been developed
2 by 5 different CDI consultants over a span of 12 years, the new VP of Finance
3 wanted to implement JATA believing they would “recoup beaucoup dollars in the
4 process.”
5 46. The effectiveness of JATA’s software in increasing revenue is driven
6 by issuing its own leading queries. JATA documents show how its software will
7 even directly prompt doctors to code specific high revenue diagnoses—e.g., hepatic
8 encephalopathy—even though CDIs are precluded from telling doctors how to
9 document care since those are considered as leading queries.
10 6. JATA’s upcoding strategies extend beyond Providence
11 47. Finally, JATA’s role in encouraging and enabling Providence to code
12 CCs and MCCs excessively was not an isolated occurrence but rather a deliberate
13 systematic strategy. Through its multifaceted investigation, Relator uncovered that
14 JATA also provided significant encouragement, training and software to assist many
15 other hospitals to submit false and inflated claims to Medicare.
16 48. Relator’s statistical analysis shows how JATA systematically
17 influenced other hospital systems to upcode MCCs, and confirms that when a
18 system switches to JATA as its CDI consultant, its usage of encephalopathy,
19 respiratory failure, and severe malnutrition (the exact same codes which Providence
20 uses to upcode) jumps significantly. This confirms statistically that JATA plays a
21 significant role in the upcoding and validates Relator’s investigation related to
22 JATA’s direct involvement in enabling hospitals to upcode. The result of this
23 statistical analysis is demonstrated in Figure 2 below. The use of encephalopathy,
24 respiratory failure, and severe malnutrition at Valley Medical Center in Renton,
25 Washington, more than doubled after they started using JATA.
26
27
28
394151.1 16 Case No. 2:17-cv-01694-PSG-SS
SECOND AMENDED COMPLAINT
Case 2:17-cv-01694-PSG-SS Document 38 Filed 08/10/18 Page 20 of 100 Page ID #:661
1 Figure 2: Rate of Encephalopathy, Respiratory Failure, and Severe Malnutrition at Valley Medical
Center Jumps Significantly After They Start Using JATA
2
3
4
5
6
7
8
9
10
11
12
13
14 C. Relator’s Methodology
15 49. Relator uncovered Providence and JATA’s fraud by employing unique
16 algorithms and statistical processes to analyze inpatient claims data for short term
17 acute care hospitals from 2011 through June 2017,7 obtained from the CMS. These
18 proprietary methods have allowed Relator to identify with specificity the false
19 claims made by Providence (with the help of JATA) to fraudulently inflate revenue
20 on Medicare claims. Relator’s analysis focused on identifying certain secondary
21 diagnoses codes—MCCs—that were fraudulently added by Providence to Medicare
22 claims to increase reimbursements.
23 50. Relator first formed groupings corresponding to 312 specific principal
24 diagnosis codes. To control for the patient’s principal diagnosis, Relator used these
25 groupings as comparative “bins.” Within each bin, Relator compared the usage rate
26
27 7
Only claims through the second quarter of 2017 were analyzed by Relator.
28 Claims after June 30, 2017, have not yet been made available to Relator.
394151.1 17 Case No. 2:17-cv-01694-PSG-SS
SECOND AMENDED COMPLAINT
Case 2:17-cv-01694-PSG-SS Document 38 Filed 08/10/18 Page 21 of 100 Page ID #:662
1 of specific MCCs at hospitals in the Providence system to usage rates in other acute
2 care inpatient hospitals. In addition, to ensure that only the truly fraudulent claims
3 were analyzed, Relator excluded any claims for which adding an MCC did not
4 increase the value.8 Similarly, Relator excluded any claims involving patients who
5 died in the course of their treatment, as these claims tend to involve patients that are
6 sicker and have higher rates of MCCs.
7 51. Given that some natural variation in usage rates among hospitals is
8 expected, Relator used two filters to further ensure that it identified truly abnormal
9 usage. First, only instances where MCCs were used more than twice the national
10 rate or were used at a rate three percentage points higher than in the other hospitals
11 were considered false claims. Second, Relator validated the results of its analysis by
12 determining the statistical significance of each fraudulent pattern used by
13 Providence. Relator only flagged claim groupings where there was less than a one-
14 thousandth percent chance of Relator’s findings being due to chance. Under this
15 approach, Relator identified 271 combinations of principal diagnosis codes and
16 Misstated MCCs in which Providence excessively upcodes. Relator included only
17 the principal diagnosis code groups that were used excessively by Providence.
18 52. For example, Providence and other hospitals have a large number of
19 claims involving a Fracture of the Neck of the Femur (hip). Relator has found that
20 among Providence’s more than 11,000 claims involving a femoral neck fracture,
21 1,429 had had an accompanying secondary MCC of encephalopathy,9 representing
22 12.34 percent of their femoral neck fracture claims. The other non-Providence
23 hospitals, used by Relator for benchmarking, had more than 1,100,000 femoral neck
24
25
8
Some diagnosis related groups do not have an MCC severity level, and as such,
26 adding an MCC does not increase the reimbursement amount.
27 9
See section IV.D.1(a) for a description of encephalopathy and the relevant codes
28 that are included.
394151.1 18 Case No. 2:17-cv-01694-PSG-SS
SECOND AMENDED COMPLAINT
Case 2:17-cv-01694-PSG-SS Document 38 Filed 08/10/18 Page 22 of 100 Page ID #:663
1 fracture claims, but only 4.46 percent of those claims reported encephalopathy as an
2 MCC. In other words, Providence coded encephalopathy on these claims at a rate
3 that is 2.77 times higher than comparable hospitals—and profited nearly $7,500
4 each time it did so.
5 53. While Relator’s precise benchmarking of medical billing is unique,
6 experts have developed and applied similar benchmarks in financial return
7 literature.10 Benchmarking has the advantage of allowing for very specific and
8 comparative groupings. This avoids imposing specific linearity on the data, which
9 in turn gives Relator’s methodology more statistical power and precision.
10 54. To further validate its conclusions and control for other explanations,
11 Relator ran a bin-based fixed effect linear regression model. Separate regressions
12 were run for claims under each principal diagnosis bin and Relator included
13 variables to control for patient characteristics such as age, gender, and race, as well
14 as county demographic factors such as the unemployment rate, median income, and
15 urban-rural differences. Additionally, variables for the length of stay and discharge
16 status were included to control for the patient’s health and overall claim severity.
17 Relator also tested for the potential impact that doctors, individual patients, and the
18 hospital’s region could have on MCC rates. Even when considering all of these
19 factors, Providence’s MCC usage rate is significantly higher than at other hospitals.
20
21
22
23
10
24 See the widely-used methodology developed by Kent Daniel, Mark Grinblatt,
Sheridan Titman, Russ Wermers, Measuring Mutual Fund Performance with
25
Characteristic‐Based Benchmarks, The Journal of Finance, vol. 52(3) (1997), at
26 1035–1058. This methodology is first applied to measuring hedge-fund
performance by John M. Griffin and Jin Xu, How Smart Are the Smart Guys? A
27
Unique View from Hedge Fund Stock Holdings, Review of Financial Studies, Vol.
28 22.7 (2009), at 2531–2570.
394151.1 19 Case No. 2:17-cv-01694-PSG-SS
SECOND AMENDED COMPLAINT
Case 2:17-cv-01694-PSG-SS Document 38 Filed 08/10/18 Page 23 of 100 Page ID #:664
1 Figure 3. Rate of Misstated MCC Upcoding by Year for Providence Versus Other Hospitals.
This figure shows the rate at which Providence is using one of the Misstated MCCs relative to other
2 hospitals over time. This analysis is based on the principal diagnosis codes listed in each section for the
specific fraudulent patterns.
3
4
5
6
7
8
9
10
11
12
13
14
15 57. The following figure shows just how drastic some of these patterns are
16 when individual hospitals are analyzed and demonstrate how widespread this abuse
17 is at nearly all the Providence hospitals. As shown in Figure 4, Providence has 14 of
18 the top 250 hospitals (out of more than 3,000 hospitals) based on the rate of
19 Misstated MCCs. If this distribution were random, the statistical probability of one
20 system having 14 of the top 250 hospitals is less than 1 in 100 million.
21
22
23
24
25
26
27
28
394151.1 21 Case No. 2:17-cv-01694-PSG-SS
SECOND AMENDED COMPLAINT
Case 2:17-cv-01694-PSG-SS Document 38 Filed 08/10/18 Page 25 of 100 Page ID #:666
4
5
6
7
8
9
10
11
12
13
14
15 58. Figure 5 below shows that Providence used a higher rate of Misstated
16 MCC codes not just in the principal diagnosis categories analyzed by Relator in this
17 complaint, but across a large variety of principal diagnosis codes. Specifically,
18 Figure 5 shows a dot for each principal diagnosis category, with the rate of
19 Misstated MCC at Providence on the x-axis and the rate of Misstated MCC at other
20 non-Providence hospitals on the y-axis. Dots to the right of the 45-degree line
21 indicate a higher rate of Misstated MCCs at Providence within that principal
22 diagnosis category than at other non-Providence hospitals. As the figure shows,
23 Providence has higher rates of Misstated MCCs across 296 of 312 (94.87%)
24 principal diagnosis categories. This demonstrates that the extent to which
25 Providence excessively upcoded on the categories identified by Relator was not
26 offset by a relative downcoding for other principal diagnosis categories as
27 Providence consistently upcodes relative to other hospitals across a large variety of
28 principal diagnosis codes.
394151.1 22 Case No. 2:17-cv-01694-PSG-SS
SECOND AMENDED COMPLAINT
Case 2:17-cv-01694-PSG-SS Document 38 Filed 08/10/18 Page 26 of 100 Page ID #:667
1
2 Figure 5. Rate of Misstated MCCs by Principal Diagnosis Code at Providence Versus Other
Hospitals.
3 For the 312 principal diagnoses with at least 100 claims at Providence (each represented by a dot), this
figure compares the rate of Misstated MCCs at Providence versus non-Providence hospitals. Red dots to the
4 right of the 45-degree line indicate Providence is coding the Misstated MCCs higher than average.
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24 (a) Encephalopathy
25 59. The first Misstated MCC fraudulently used by Providence to make
26 false claims is encephalopathy. The codes included with encephalopathy are listed
27 in Table 1. Encephalopathy is a term for brain disease or damage to the brain where
28 the brain is regarded as “altered in its structure or function.” The telltale symptom is
394151.1 23 Case No. 2:17-cv-01694-PSG-SS
SECOND AMENDED COMPLAINT
Case 2:17-cv-01694-PSG-SS Document 38 Filed 08/10/18 Page 27 of 100 Page ID #:668
1 an altered mental state, but altered mental state alone is insufficient for diagnosing
2 encephalopathy. Encephalopathy can be acute or chronic, so the related signs and
3 symptoms can be varied as well. This condition commonly manifests as confusion,
4 agitation, or lethargy, but may include aphasia (altered speech), ataxia (altered gait)
5 and memory loss.
6
7 Table 1. List of Encephalopathy ICD-9 and ICD-10 Diagnosis Codes.
ICD-9 Diagnosis Code Description
8 34830 Encephalopathy, unspecified
34831 Metabolic encephalopathy
9 34839 Other encephalopathy
34982 Toxic encephalopathy
10 ICD-10 Diagnosis Code Description
G92 Toxic encephalopathy
11 G9340 Encephalopathy, unspecified
G9341 Metabolic encephalopathy
12 G9349 Other encephalopathy
I6783 Posterior reversible encephalopathy syndrome
13
14 60. The most common causes of encephalopathy are liver damage, cerebral
15 anoxia (severe lack of oxygen to the brain) or kidney failure. Because the causes are
16 extremely varied, no single lab test can prove the presence of encephalopathy.
17 Therefore, in diagnosing the condition, a medical practitioner must keep multiple
18 considerations in mind. The challenge is to properly identify the root cause of the
19 symptoms observed and eliminate unlikely causes based on objective signs.
20 61. Encephalopathy is distinguishable from conditions that have similar
21 symptoms. In elderly hospital patients, for instance, temporary instances of
22 lethargy, agitation and confusion are commonly observed, often right after an
23 intense surgery or as the result of a urinary tract infection. These same signs can be
24 observed in patients as “sundowning” or “late-day confusion” in the late afternoon
25 or evening, but these effects are temporary and are actually related to chronic
26 dementia, not encephalopathy.
27 62. Between 2011 and June 2017, Providence was 2.07 times more likely
28 to code encephalopathy than other hospitals. During this period, Providence coded
394151.1 24 Case No. 2:17-cv-01694-PSG-SS
SECOND AMENDED COMPLAINT
Case 2:17-cv-01694-PSG-SS Document 38 Filed 08/10/18 Page 28 of 100 Page ID #:669
7
8
9
10
11
12
13
14
15
16
17
18 (i) Specific Patterns of Fraud with Encephalopathy
19 63. Table 2 provides a list of the principal diagnosis codes used by
20 Providence to upcode with encephalopathy. Relator identified 215 principal
21 diagnosis codes in conjunction with which Providence coded encephalopathy at a
22 rate at least two times and/or three percentage points higher than the nationwide
23 average. Relator has included only patterns that were statistically significant at the
24 99.9% level, meaning it is virtually impossible the patterns are due to chance.
25
26
27
28
394151.1 25 Case No. 2:17-cv-01694-PSG-SS
SECOND AMENDED COMPLAINT
Case 2:17-cv-01694-PSG-SS Document 38 Filed 08/10/18 Page 29 of 100 Page ID #:670
1 Providence Num. of
% with MCC Rate Relative to Fraud
2 in Other % with MCC Nationwide Claims at
Principal Diagnosis Hospitals in Providence Average Providence
3 Fracture of Vertebral Column
without Mention of Spinal Cord 4.03% 10.32% 256% 153
Injury
4 Secondary Malignancy of
10.08% 21.88% 217% 135
Brain/spine
5 Other Connective Tissue Disease 4.68% 10.43% 223% 135
Postoperative Infection 4.09% 8.22% 201% 134
6 Alcohol-related Disorders 6.86% 10.90% 159% 127
Hemorrhage from Gastrointestinal
7 Ulcer
2.96% 6.74% 228% 117
Other Intestinal Obstruction 1.50% 3.67% 245% 113
8 Cancer of Bronchus; Lung 2.84% 6.26% 220% 112
Other Fluid and Electrolyte
9 Disorders
8.35% 18.55% 222% 111
Pulmonary Heart Disease 2.41% 4.95% 206% 101
10 Acute Pancreatitis 2.14% 5.63% 263% 99
Infections of Kidney 4.82% 11.71% 243% 96
11 Other and Unspecified
2.01% 5.02% 249% 91
Gastrointestinal Disorders
12 Other Fracture of Lower Limb 3.35% 9.21% 275% 88
Fracture of Pelvis 2.81% 7.74% 275% 82
13 Congestive Heart Failure 1.71% 4.92% 288% 80
Other Nervous System Symptoms
14 and Disorders
3.08% 8.41% 273% 79
Cancer of Brain and Nervous
15 System
9.78% 21.01% 215% 75
Non-Hodgkin’s Lymphoma 6.32% 14.57% 231% 75
16 Delirium Dementia and Amnestic
14.98% 27.17% 181% 74
and Other Cognitive Disorders
17 Other and Unspecified Hereditary
and Degenerative Nervous 8.05% 13.95% 173% 73
18 Conditions
Disorders of Mineral Metabolism 12.55% 24.83% 198% 73
19 Cancer of Colon 2.20% 5.26% 239% 72
Other Injuries and Conditions Due
20 6.10% 13.76% 226% 71
to External Causes
Fracture of Ankle 2.56% 7.47% 292% 70
21 Fracture of Humerus 3.24% 8.22% 254% 65
Other Peripheral and Visceral
2.44% 6.16% 252% 65
22 Atherosclerosis
Other Cardiac Dysrhythmias 2.44% 5.64% 231% 65
23 Infective Arthritis and
Osteomyelitis (except That Caused 3.36% 7.42% 221% 62
24 by Tb or STD)
Diverticulitis 1.36% 3.09% 227% 59
25 Peritoneal or Intestinal Adhesions 2.42% 4.85% 200% 59
Other Bacterial Pneumonia 7.13% 10.59% 149% 57
26 Intervertebral Disc Disorders 1.45% 2.94% 203% 57
Other Specified Septicemia 16.60% 26.52% 160% 57
27 Other and Unspecified Viral
7.59% 16.46% 217% 56
Infection
28 Spinal Stenosis; Lumbar Region 1.66% 3.38% 203% 56
1 Providence Num. of
% with MCC Rate Relative to Fraud
2 in Other % with MCC Nationwide Claims at
Principal Diagnosis Hospitals in Providence Average Providence
3 Gangrene 4.75% 10.45% 220% 56
Encephalitis (except That Caused
25.33% 55.49% 219% 55
by Tb or STD)
4 Other Aneurysm 3.79% 8.62% 228% 54
Other Venous Embolism and
5 Thrombosis
1.60% 4.05% 253% 54
Pleurisy; Pleural Effusion 2.42% 5.92% 245% 53
6 Liver Abscess and Sequelae of
6.26% 9.84% 157% 53
Chronic Liver Disease
7 Atherosclerosis of Arteries of
1.55% 3.63% 234% 52
Extremities
8 Hypotension 4.81% 12.48% 260% 51
Cancer of Pancreas 2.68% 7.75% 289% 50
9 Calculus of Bile Duct 1.82% 4.35% 239% 49
Melena 1.96% 5.11% 260% 49
10 Fracture of Ribs; Closed 2.89% 8.45% 292% 49
Noninfectious Gastroenteritis 1.66% 5.22% 315% 48
11 Crushing Injury or Internal Injury 3.56% 10.35% 291% 48
Sinoatrial Node Dysfunction 2.41% 4.84% 201% 48
12 Hyperpotassemia 4.57% 8.26% 181% 47
Superficial Injury; Contusion 2.78% 8.42% 303% 46
13 Suicide and Intentional Self-
26.39% 44.19% 167% 46
inflicted Injury
14 Chronic Rheumatic Disease of the
3.31% 7.35% 222% 46
Heart Valves
15 Other Pneumonia 3.92% 10.22% 261% 45
Diabetes with Circulatory
16 Manifestations
3.86% 8.70% 225% 44
Late Effects of Cerebrovascular
17 Disease
10.25% 24.50% 239% 43
Secondary Malignancy of Bone 4.42% 10.85% 246% 42
18 Osteoarthritis; Generalized and
0.65% 1.67% 256% 41
Unspecified
19 Other and Unspecified Metabolic;
Nutritional; and Endocrine 5.72% 14.14% 247% 40
20 Disorders
Other Disorders of Stomach and
21 1.84% 4.25% 230% 40
Duodenum
Acute Posthemorrhagic Anemia 2.70% 6.72% 249% 40
22 Other Diseases of the Nervous
7.84% 14.67% 187% 38
System and Sense Organs
23 Cholelithiasis with Acute
1.81% 3.64% 201% 38
Cholecystitis
24 Diseases of White Blood Cells 2.63% 6.46% 246% 36
Hemorrhage or Hematoma
1.49% 4.00% 268% 36
25 Complicating A Procedure
Diverticulosis 1.34% 2.84% 213% 34
26 Other Biliary Tract Disease 2.14% 5.88% 275% 34
Other Bone Disease and
1.57% 4.08% 259% 33
27 Musculoskeletal Deformities
Complications of Transplants and
2.33% 6.46% 278% 33
28 Reattached Limbs
1 Providence Num. of
% with MCC Rate Relative to Fraud
2 in Other % with MCC Nationwide Claims at
Principal Diagnosis Hospitals in Providence Average Providence
3 Cancer of Liver and Intrahepatic
3.91% 11.11% 285% 32
Bile Duct
Malaise and Fatigue 2.98% 8.46% 284% 32
4 Concussion 6.85% 24.31% 355% 32
Other Mycoses 11.55% 19.54% 169% 31
5 Anemia; Unspecified 1.66% 5.47% 331% 31
Disorders of the Peripheral
6 Nervous System
2.82% 7.21% 256% 30
Neoplasms of Unspecified Nature
7 or Uncertain Behavior
2.84% 6.25% 220% 30
Empyema and Pneumothorax 2.86% 7.32% 256% 30
8 Other Non-Traumatic Joint
1.77% 5.28% 297% 30
Disorders
9 Other Diseases of the Respiratory
2.23% 9.24% 415% 30
System
10 Other Esophageal Disorders 1.64% 3.70% 226% 30
Other Cellulitis and Abscess 2.04% 4.92% 241% 29
11 Essential Hypertension 2.23% 6.97% 313% 29
Fracture of Tibia and Fibula 2.59% 5.98% 231% 28
12 Decubitus Ulcer 4.21% 9.09% 216% 27
Constipation 1.82% 5.44% 299% 27
13 Gout and Other Crystal
2.53% 9.61% 380% 27
Arthropathies
14 Multiple Myeloma 9.31% 17.30% 186% 27
Other and Ill-defined
15 Cerebrovascular Disease
5.78% 10.11% 175% 27
Cholecystitis without
16 Cholelithiasis
2.23% 5.12% 229% 27
Cancer of Stomach 2.25% 7.60% 338% 26
17 Other Hypertensive Complications 4.80% 9.94% 207% 26
Other and Unspecified Liver
18 Disorders
6.56% 12.71% 194% 26
Hematemesis 3.17% 8.11% 255% 26
19 Hematuria 1.80% 7.80% 434% 25
Cancer of Head and Neck 2.29% 9.91% 433% 25
20 Leukemias 4.15% 8.79% 212% 25
Cardiac Arrest and Ventricular
21 8.42% 16.50% 196% 25
Fibrillation
Cellulitis and Abscess of Arm 2.12% 4.85% 229% 25
22 Herpes Zoster Infection 7.39% 16.36% 221% 25
Fracture of Radius and Ulna 2.25% 5.87% 260% 25
23 Cellulitis and Abscess of Foot 1.88% 6.34% 337% 25
Urinary Complications 6.98% 16.29% 233% 25
24 Benign Neoplasm of Cerebral
7.42% 12.61% 170% 24
Meninges
25 Chronic Obstructive Asthma with
0.86% 2.65% 307% 24
Acute Exacerbation
26 Atrial Flutter 1.11% 2.74% 247% 23
Other Injury and Poisoning 3.28% 15.85% 484% 23
27 Respiratory Failure 14.70% 24.36% 166% 23
28
394151.1 29 Case No. 2:17-cv-01694-PSG-SS
SECOND AMENDED COMPLAINT
Case 2:17-cv-01694-PSG-SS Document 38 Filed 08/10/18 Page 33 of 100 Page ID #:674
1 Providence Num. of
% with MCC Rate Relative to Fraud
2 in Other % with MCC Nationwide Claims at
Principal Diagnosis Hospitals in Providence Average Providence
3 Acute Upper Respiratory
Infections of Multiple or 3.27% 13.36% 409% 22
Unspecified Sites
4 Cancer of Bladder 2.71% 5.74% 212% 22
Diabetes with Renal
5 Manifestations
6.63% 14.83% 224% 22
Other Cns Infection and
6 Poliomyelitis
13.11% 23.76% 181% 22
Other Endocrine; Nutritional; and
7 Metabolic Diseases and Immunity 3.52% 11.40% 324% 21
Disorders
8 Abdominal Aortic Aneurysm;
1.23% 2.63% 213% 21
without Rupture
9 Paralytic Ileus 2.63% 5.43% 207% 21
Hepatitis 4.55% 8.08% 178% 21
10 Other and Unspecified Circulatory
3.46% 7.72% 223% 20
Disease
11 Other Diseases of the Circulatory
1.94% 5.02% 259% 20
System
12 Multiple Sclerosis 3.27% 9.84% 301% 20
Other Specified Anemia 1.99% 6.04% 304% 19
13 Esophagitis 1.72% 5.11% 297% 19
Chronic Kidney Disease 6.74% 12.39% 184% 19
14 Cancer of Rectum and Anus 1.99% 4.10% 206% 19
Incisional Hernia with
15 1.59% 4.27% 269% 19
Obstruction/gangrene
Calculus of Ureter 1.22% 3.09% 253% 19
16 Peritonitis and Intestinal Abscess 3.42% 6.94% 203% 18
Other and Unspecified Lower
17 1.78% 4.48% 252% 18
Respiratory Disease
Poisoning by Nonmedicinal
18 14.16% 30.56% 216% 18
Substances
Abdominal Pain 0.91% 2.17% 239% 17
19 Inflammatory Conditions of Male
2.53% 9.16% 362% 17
Genital Organs
20 Shock 19.47% 34.82% 179% 17
Secondary Malignancy of Liver 3.87% 7.35% 190% 17
21 Arterial Embolism and
Thrombosis of Lower Extremity 3.23% 6.27% 194% 16
22 Artery
Other Back Pain and Disorders 1.46% 2.92% 200% 16
23 Hemorrhage of Rectum and Anus 1.76% 5.21% 295% 16
Cancer of Other GI Organs;
2.28% 5.12% 224% 16
24 Peritoneum
Cirrhosis of Liver without Mention
4.51% 7.91% 175% 15
25 of Alcohol
Migraine 2.30% 7.56% 329% 15
26 Cancer of Ovary 1.58% 3.92% 248% 15
Allergic Reactions 2.74% 8.37% 306% 15
27 Other Infectious and Parasitic
6.47% 12.95% 200% 14
Diseases
28
394151.1 30 Case No. 2:17-cv-01694-PSG-SS
SECOND AMENDED COMPLAINT
Case 2:17-cv-01694-PSG-SS Document 38 Filed 08/10/18 Page 34 of 100 Page ID #:675
1 Providence Num. of
% with MCC Rate Relative to Fraud
2 in Other % with MCC Nationwide Claims at
Principal Diagnosis Hospitals in Providence Average Providence
3 Cholelithiasis with Other
1.35% 3.63% 268% 14
Cholecystitis
Cardiac Complications 2.73% 8.33% 305% 14
4 Unspecified Gastritis and
1.91% 4.55% 238% 14
Gastroduodenitis
5 Unstable Angina (intermediate
1.20% 2.45% 205% 14
Coronary Syndrome)
6 Inguinal Hernia with Obstruction
2.20% 4.84% 220% 14
or Gangrene
7 Cellulitis and Abscess of Hand 1.49% 4.88% 327% 13
Obesity 0.71% 3.15% 447% 13
8 Open Wounds of Extremities 1.71% 7.59% 443% 13
Lumbago 2.43% 6.83% 280% 13
9 Diseases of Mouth; Excluding
3.45% 8.82% 256% 13
Dental
10 Other Arterial Embolism and
2.52% 5.84% 232% 13
Thrombosis
11 Cancer of Kidney and Renal Pelvis 4.01% 15.18% 378% 13
Coma; Stupor; and Brain Damage 7.54% 14.88% 197% 12
12 Open Wounds of Head; Neck; and
2.88% 9.09% 316% 12
Trunk
13 Other and Unspecified
3.88% 12.00% 309% 12
Genitourinary Symptoms
14 Backache; Unspecified 2.53% 10.13% 401% 12
Gastric Ulcer 2.33% 6.29% 270% 12
15 Other Upper Respiratory Disease 1.63% 4.35% 266% 12
Dysphagia 3.20% 7.78% 243% 12
16 Rheumatoid Arthritis and Related
1.64% 5.82% 355% 11
Disease
17 Other Diseases of the Blood and
2.73% 8.17% 299% 11
Blood-forming Organs
18 Calculus of Kidney 1.23% 4.19% 342% 11
Other Diseases of the
19 2.50% 6.07% 243% 11
Genitourinary System
Cellulitis and Abscess of Fingers
20 1.45% 4.72% 326% 11
and Toes
Other Diseases of Veins and
1.51% 4.65% 308% 11
21 Lymphatics
22
64. The following figure shows just how drastic some of these patterns are
23
when individual hospitals are analyzed, and starkly demonstrate Providence’s
24
widespread abuse. As shown in Figure 7, Providence hospitals are consistently
25
among those with the highest rates of encephalopathy. In fact, Providence has 18 of
26
the top 250 hospitals in the country (out of more than 3,000 total hospitals) based on
27
28
394151.1 31 Case No. 2:17-cv-01694-PSG-SS
SECOND AMENDED COMPLAINT
Case 2:17-cv-01694-PSG-SS Document 38 Filed 08/10/18 Page 35 of 100 Page ID #:676
6
7
8
9
10
11
12
13
14
15
16
17
18 65. Figure 8 below shows that Providence not only used a higher rate of
19 encephalopathy in the few categories listed above, but also across a large variety of
20 principal diagnosis codes. The red dots to the right of the 45-degree line indicate
21 higher rates of encephalopathy at Providence versus the nationwide average. This
22 figure shows that Providence has a higher rate of encephalopathy for 304 out of 312
23 (97.44%) principal diagnosis categories. In other words, the extent to which
24 Providence excessively upcoded encephalopathy on the categories listed in Table 2
25 above was not offset by relative downcoding in other principal diagnosis categories.
26
27
28
394151.1 32 Case No. 2:17-cv-01694-PSG-SS
SECOND AMENDED COMPLAINT
Case 2:17-cv-01694-PSG-SS Document 38 Filed 08/10/18 Page 36 of 100 Page ID #:677
1 Figure 8. Rate of Encephalopathy by Principal Diagnosis Code at Providence Versus Other Hospitals.
For the 312 principal diagnoses with at least 100 claims at Providence (each represented by a dot), this
2 figure compares the rate of encephalopathy at Providence versus non-Providence hospitals. Red dots to the
right of the 45-degree line indicate Providence is coding encephalopathy at a rate higher than the
3 nationwide average.
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24 (ii) Specific False Claims with Encephalopathy
25 66. The Relator has identified many specific false Medicare claims
26 submitted by Providence involving encephalopathy. The following table includes
27 50 examples.
28
394151.1 33 Case No. 2:17-cv-01694-PSG-SS
SECOND AMENDED COMPLAINT
Case 2:17-cv-01694-PSG-SS Document 38 Filed 08/10/18 Page 37 of 100 Page ID #:678
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
394151.1 34 Case No. 2:17-cv-01694-PSG-SS
SECOND AMENDED COMPLAINT
Case 2:17-cv-01694-PSG-SS Document 38 Filed 08/10/18 Page 38 of 100 Page ID #:679
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
394151.1 35 Case No. 2:17-cv-01694-PSG-SS
SECOND AMENDED COMPLAINT
Case 2:17-cv-01694-PSG-SS Document 38 Filed 08/10/18 Page 39 of 100 Page ID #:680
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
394151.1 36 Case No. 2:17-cv-01694-PSG-SS
SECOND AMENDED COMPLAINT
Case 2:17-cv-01694-PSG-SS Document 38 Filed 08/10/18 Page 40 of 100 Page ID #:681
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
394151.1 37 Case No. 2:17-cv-01694-PSG-SS
SECOND AMENDED COMPLAINT
Case 2:17-cv-01694-PSG-SS Document 38 Filed 08/10/18 Page 41 of 100 Page ID #:682
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
(b) Respiratory Failure
22
67. The second Misstated MCC that Providence used at an excessive rate is
23
respiratory failure, which includes pulmonary insufficiency. The codes classified as
24
respiratory failure are listed in Table 4 below. Respiratory failure is a syndrome
25
characterized by poor gas transfer in the lungs at the alveolar and capillary levels as
26
a result of a problem making it difficult to breathe. It can be acute or chronic.
27
There are two types: the first and most common is hypoxemia (“oxygenation
28
394151.1 38 Case No. 2:17-cv-01694-PSG-SS
SECOND AMENDED COMPLAINT
Case 2:17-cv-01694-PSG-SS Document 38 Filed 08/10/18 Page 42 of 100 Page ID #:683
1 failure”), and the second type demonstrates both hypoxemia and hypercapnia
2 (“ventilatory failure”). Respiratory failure can be acute and life-threatening, or
3 chronic and manageable with modifications.
4
5 Table 4. List of Respiratory Failure ICD-9 and ICD-10 Diagnosis Codes.
6 ICD-9
Diagnosis Code Description
7 5184 Acute edema of lung, unspecified
5185 Pulmonary insufficiency following trauma and surgery
8 51851 Acute respiratory failure following trauma and surgery
51852 Other pulmonary insufficiency not elsewhere classified following trauma/surgery
51853 Acute and chronic respiratory failure following trauma and surgery
9 51881 Acute respiratory failure
51884 Acute and chronic respiratory failure
10
ICD-10
11 Diagnosis Code Description
J810 Acute pulmonary edema
12 J951 Acute pulmonary insufficiency following thoracic surgery
J952 Acute pulmonary insufficiency following nonthoracic surgery
13 J953 Chronic pulmonary insufficiency following surgery
J95821 Acute postprocedural respiratory failure
14 J95822 Acute and chronic postprocedural respiratory failure
J9600 Acute respiratory failure, unspecified whether with hypoxia or hypercapnia
15 J9601 Acute respiratory failure with hypoxia
J9602 Acute respiratory failure with hypercapnia
16 J9620 Acute and chronic respiratory failure, unspecified whether with hypoxia or hypercapnia
J9621 Acute and chronic respiratory failure with hypoxia
17 J9622 Acute and chronic respiratory failure with hypercapnia
J9690 Respiratory failure, unspecified, unspecified whether with hypoxia or hypercapnia
18 J9691 Respiratory failure, unspecified with hypoxia
J9692 Respiratory failure, unspecified with hypercapnia
19
20 68. The possible root causes are myriad, and may include poor circulation,
21 neuromuscular disease, chronic bronchitis, COPD, obesity or drug use, an
22 obstructing object, or an injury to the brain or spinal cord. The signs and symptoms
23 are bluish skin, shortness of breath, labored breathing and feeling unable to get
24 enough air. The patient may also become very sleepy, lose consciousness, be
25 confused, or have arrhythmia. After listening to the patient’s heartbeat and lungs, a
26 pulse oximetry test, an arterial blood gas test from a blood draw, and a chest x-ray
27 can together help determine a proper diagnosis.
28
394151.1 39 Case No. 2:17-cv-01694-PSG-SS
SECOND AMENDED COMPLAINT
Case 2:17-cv-01694-PSG-SS Document 38 Filed 08/10/18 Page 43 of 100 Page ID #:684
16
17
18
19
20
21
22
23
24
25
26
27
28
394151.1 40 Case No. 2:17-cv-01694-PSG-SS
SECOND AMENDED COMPLAINT
Case 2:17-cv-01694-PSG-SS Document 38 Filed 08/10/18 Page 44 of 100 Page ID #:685
1 % with MCC
Providence Num. of
% with MCC Rate Relative Fraud
Principal Diagnosis in Other
2 in Providence to Other Claims at
Hospitals
Hospitals Providence
3 Leukemias 7.52% 12.45% 166% 27
Allergic Reactions 8.81% 14.83% 168% 16
4 Other Infectious and Parasitic Diseases 6.80% 13.84% 203% 16
5
6 72. Figure 10 shows the extent of the fraud at Providence hospitals. The
7 figure shows the rate of respiratory failure for claims in the above principal
8 diagnosis categories. Providence had 9 of the top 250 hospitals based on respiratory
9 failure usage rate for the above principal diagnosis categories. The probability of
10 Providence randomly having 9 of its hospitals in the top 250 hospitals is less than 1
11 in 30 thousand.
12
13 Figure 10. Rate of Respiratory Failure by Hospital.
The following figure plots every Provider with at least 100 claims in the above principal diagnosis code
14 categories. 3,173 hospitals had at least 100 claims with the relevant principal diagnosis codes listed in Table
5. The hospitals are plotted along the x-axis in order of percentage of respiratory failure usage. The
15 Providence hospitals are in red, while other hospitals are in blue.
16
17
18
19
20
21
22
23
24
25
26
27
28
394151.1 42 Case No. 2:17-cv-01694-PSG-SS
SECOND AMENDED COMPLAINT
Case 2:17-cv-01694-PSG-SS Document 38 Filed 08/10/18 Page 46 of 100 Page ID #:687
1 73. Figure 11 below shows that Providence used a higher rate of respiratory
2 failure not just in the few categories listed above, but across a large variety of
3 principal diagnosis codes. The red dots to the right of the 45-degree line indicate
4 higher rates of respiratory failure at Providence versus the nationwide average. This
5 figure shows that Providence has a higher rate of respiratory failure for 230 out of
6 312 (73.72%) principal diagnosis categories. In other words, the extent to which
7 Providence excessively upcoded respiratory failure on the categories listed in Table
8 5 above was not offset by relative downcoding in other principal diagnosis
9 categories.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
394151.1 43 Case No. 2:17-cv-01694-PSG-SS
SECOND AMENDED COMPLAINT
Case 2:17-cv-01694-PSG-SS Document 38 Filed 08/10/18 Page 47 of 100 Page ID #:688
1 Figure 11. Rate of Respiratory Failure by Principal Diagnosis Code at Providence Versus Other
Hospitals.
2 For the 312 principal diagnoses with at least 100 claims at Providence (each represented by a dot), this
figure compares the rate of respiratory failure at Providence versus non-Providence hospitals. Red dots to
3 the right of the 45-degree line indicate Providence is coding respiratory failure at a rate higher than the
nationwide average.
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24 (ii) Specific False Claims with Respiratory Failure
25 74. The Relator has identified many specific false Medicare claims
26 submitted by Providence involving respiratory failure. The following table includes
27 50 examples.
28
394151.1 44 Case No. 2:17-cv-01694-PSG-SS
SECOND AMENDED COMPLAINT
Case 2:17-cv-01694-PSG-SS Document 38 Filed 08/10/18 Page 48 of 100 Page ID #:689
1 Beneficiary
Discharge
Age/
Principal MCC False
DRG w/ False
ID / Claim Hospital Gender/ False Claim
Date Diagnosis Claim
2 ID Race Claim Amount
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
394151.1 46 Case No. 2:17-cv-01694-PSG-SS
SECOND AMENDED COMPLAINT
Case 2:17-cv-01694-PSG-SS Document 38 Filed 08/10/18 Page 50 of 100 Page ID #:691
1 Beneficiary
Discharge
Age/
Principal MCC False
DRG w/ False
ID / Claim Hospital Gender/ False Claim
Date Diagnosis Claim
2 ID Race Claim Amount
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
394151.1 47 Case No. 2:17-cv-01694-PSG-SS
SECOND AMENDED COMPLAINT
Case 2:17-cv-01694-PSG-SS Document 38 Filed 08/10/18 Page 51 of 100 Page ID #:692
1 Beneficiary
Discharge
Age/
Principal MCC False
DRG w/ False
ID / Claim Hospital Gender/ False Claim
Date Diagnosis Claim
2 ID Race Claim Amount
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
394151.1 48 Case No. 2:17-cv-01694-PSG-SS
SECOND AMENDED COMPLAINT
Case 2:17-cv-01694-PSG-SS Document 38 Filed 08/10/18 Page 52 of 100 Page ID #:693
1 Beneficiary
Discharge
Age/
Principal MCC False
DRG w/ False
ID / Claim Hospital Gender/ False Claim
Date Diagnosis Claim
2 ID Race Claim Amount
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
394151.1 49 Case No. 2:17-cv-01694-PSG-SS
SECOND AMENDED COMPLAINT
Case 2:17-cv-01694-PSG-SS Document 38 Filed 08/10/18 Page 53 of 100 Page ID #:694
1 Beneficiary
Discharge
Age/
Principal MCC False
DRG w/ False
ID / Claim Hospital Gender/ False Claim
Date Diagnosis Claim
2 ID Race Claim Amount
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19 (c) Severe Malnutrition
20 75. The final Misstated MCC that Providence used to commit fraud at a
21 higher rate was severe malnutrition. There are three malnutrition codes, listed in
22 Table 7, that are considered MCCs. Severe malnutrition in the elderly is a disorder
23 of extreme lack of nutrition involving the highest level of protein-energy
24 malnutrition and protein-calorie malnutrition. Another rare form, Kwashiorkor
25 malnutrition, is common in Sub-Saharan Africa, and is unlikely to be present in the
26 elderly in the United States. Nutritional marasmus is caused by insufficient
27 nutrients and is most common in children. In the elderly, malnourishment can
28 manifest for a variety of reasons including anorexia, dehydration and malabsorption.
394151.1 50 Case No. 2:17-cv-01694-PSG-SS
SECOND AMENDED COMPLAINT
Case 2:17-cv-01694-PSG-SS Document 38 Filed 08/10/18 Page 54 of 100 Page ID #:695
1 Figure 12. Rate of Severe Malnutrition by Year for Providence Versus Other Hospitals.
This figure shows the rate of severe malnutrition at Providence and at other hospitals from 2011 through
2 June 2017, when added to the suspicious principal diagnosis codes listed in Table 8 below.
3
4
5
6
7
8
9
10
11
12
13
14 (i) Specific Patterns of Fraud with Severe
15 Malnutrition
16 78. The following table provides a list of the principal diagnosis codes used
17 by Providence to upcode with severe malnutrition. Relator identified 28 principal
18 diagnosis codes in conjunction with which Providence coded severe malnutrition at
19 a rate at least two times and/or at three percentage points higher than the nationwide
20 average. Only patterns that were statistically significant at the 99.9% level, meaning
21 virtually impossible to be due to chance, are included.
22
23 Table 8. Patterns Used by Providence to Upcode with Severe Malnutrition.
The following table lists the principal diagnosis categories in which Providence excessively upcodes with
24 severe malnutrition. 2 principal diagnosis categories with fewer than 11 fraudulent claims at Providence have
been omitted from the table.
25 Principal Diagnosis Code % with MCC % with MCC Providence Num. of
Code in Other Code at Rate Relative Fraud Claims
26 Hospitals Providence to Other at Providence
Hospitals
27 Obstructive Chronic Bronchitis 0.96% 2.19% 228% 103
Other Secondary Malignancy 6.76% 11.71% 173% 81
28 Hypovolemia 3.57% 6.77% 189% 77
394151.1 52 Case No. 2:17-cv-01694-PSG-SS
SECOND AMENDED COMPLAINT
Case 2:17-cv-01694-PSG-SS Document 38 Filed 08/10/18 Page 56 of 100 Page ID #:697
18 79. The following figure shows just how drastic some of these patterns are
19 when individual hospitals are analyzed, and demonstrate Providence’s widespread
20 abuse. As shown in Figure 13, Providence hospitals are consistently among the
21 highest based on rate of severe malnutrition. Specifically, Providence had 12 of the
22 top 250 hospitals with the highest rate of severe malnutrition. The probability of
23 this occurring due to random chance is less than 1 in 25 million.
24
25
26
27
28
394151.1 53 Case No. 2:17-cv-01694-PSG-SS
SECOND AMENDED COMPLAINT
Case 2:17-cv-01694-PSG-SS Document 38 Filed 08/10/18 Page 57 of 100 Page ID #:698
1 Figure 14. Rate of Severe Malnutrition by Principal Diagnosis Code at Providence Versus Other
Hospitals.
2 For the 312 principal diagnoses with at least 100 claims at Providence (each represented by a dot), this
figure compares the rate of severe malnutrition at Providence versus non-Providence hospitals. Red dots to
3 the right of the 45-degree line indicate Providence is coding severe malnutrition at a rate higher than the
nationwide average.
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24 (ii) Specific False Claims with Severe Malnutrition
25 81. The Relator has identified many specific false Medicare claims
26 submitted by Providence involving severe malnutrition. The following table
27 includes 25 examples.
28
394151.1 55 Case No. 2:17-cv-01694-PSG-SS
SECOND AMENDED COMPLAINT
Case 2:17-cv-01694-PSG-SS Document 38 Filed 08/10/18 Page 59 of 100 Page ID #:700
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
394151.1 56 Case No. 2:17-cv-01694-PSG-SS
SECOND AMENDED COMPLAINT
Case 2:17-cv-01694-PSG-SS Document 38 Filed 08/10/18 Page 60 of 100 Page ID #:701
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
394151.1 57 Case No. 2:17-cv-01694-PSG-SS
SECOND AMENDED COMPLAINT
Case 2:17-cv-01694-PSG-SS Document 38 Filed 08/10/18 Page 61 of 100 Page ID #:702
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16 2. Alternative Hypotheses for Excessive Rates of Misstated
17 MCCs Do Not Stand and Confirm that Providence
18 Fraudulently Billed Medicare
19 82. To further demonstrate the Defendants’ fraud and determine
20 responsibility for the excessively high rates of Misstated MCCs, Relator has
21 analyzed whether the statistically aberrant rates of Misstated MCCs described above
22 could be attributed to a variety of other factors. First, Relator ran a fixed effect
23 linear regression model to control for a variety of possible explanations for MCCs,
24 including patient characteristics and county demographic data. Second, Relator
25 analyzed the change in coding practices that occurred at St. Joseph Health after its
26 merger with Providence. Third, Relator considered whether the patient’s attending
27 physician is responsible for excessive Misstated MCCs by analyzing a subset of
28 claims where Providence and other hospitals shared a common physician. Fourth,
394151.1 58 Case No. 2:17-cv-01694-PSG-SS
SECOND AMENDED COMPLAINT
Case 2:17-cv-01694-PSG-SS Document 38 Filed 08/10/18 Page 62 of 100 Page ID #:703
1 Relator analyzed a subset of claims where Providence and other hospitals shared
2 common patients. Finally, Relator analyzed the upcoding rate for Providence and
3 other hospitals in the same metropolitan statistical area (“MSA”) to determine
4 whether the excessive rates of Misstated MCCs is due to regional differences. As
5 discussed further below, these analyses prove that the excessive rates of Misstated
6 MCCs can be directly attributed to the Defendants’ fraudulent activity as opposed to
7 external factors, indicating that the fraud was known by the system and was
8 intentional.
9 (a) Patient Characteristics and Demographics Do Not
10 Explain the Excessive Rates of Misstated MCCs at
11 Providence
12 83. The Relator developed a proprietary linear regression model to control
13 for the possibility that there are certain patient characteristics which might indicate a
14 higher likelihood a patient would have a MCC, allowing Relator to isolate and
15 calculate the specific impact Defendants had on the abuse of a Misstated MCC code
16 after controlling for other characteristics. These characteristics include basic patient
17 characteristics, such as the age, gender, and race, as well as characteristics relating
18 to the patient’s inpatient stay, including principal diagnosis, length of stay, and
19 discharge status. Relator also used county-level demographic data, such as
20 unemployment rate, percent of population without a high school diploma, median
21 income, and the rural-urban continuum codes from the Department of Agriculture as
22 control variables.12 These county demographic variables provided Relator with a
23 proxy for the income levels, education levels, and access to care available to each
24 patient. Regression analysis is well established and has been used to pinpoint actors
25
26
12
27 The Rural-Urban Continuum Codes measure whether each county is in a metro
or non-metro area and reflect the overall size of the metropolitan area.
28
394151.1 59 Case No. 2:17-cv-01694-PSG-SS
SECOND AMENDED COMPLAINT
Case 2:17-cv-01694-PSG-SS Document 38 Filed 08/10/18 Page 63 of 100 Page ID #:704
7 . . . . . log
8
. . . . .
9
.
10
Panel B: Explanation of Regression Variables
11
Variable Description
12
Intercept
13 Whether the claim included a MCC
Whether the patient was treated at Providence
14 Patient’s age on the claim (6 age groups)
Patient’s race on the claim (7 race categories)
15
Patient’s gender
16 The patient’s log length of stay at the hospital for claim
The patient’s discharge status
17 Season control variable for the claim (Winter, Spring, Summer, Fall)
Patient’s rural urban continuum code based on the county
18
County poverty rate in 2014
19 County unemployment rate in 2014
County log median income in 2014
20 Error term
21
22 87. By controlling for these characteristics, the regression model allowed
23 Relator to isolate the impact that being treated at a Providence hospital would have
24 on a patient’s expected likelihood of being diagnosed with one of the Misstated
25 MCCs. For example, given two patients with fracture of neck of femur, with the
26 same age and gender, from the same county, admitted during the same season, and
27 with the same length of stay, the patient treated at Providence would be 208.57%
28 times more likely to be diagnosed with encephalopathy. Figure 15 shows the results
394151.1 61 Case No. 2:17-cv-01694-PSG-SS
SECOND AMENDED COMPLAINT
Case 2:17-cv-01694-PSG-SS Document 38 Filed 08/10/18 Page 65 of 100 Page ID #:706
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
394151.1 63 Case No. 2:17-cv-01694-PSG-SS
SECOND AMENDED COMPLAINT
Case 2:17-cv-01694-PSG-SS Document 38 Filed 08/10/18 Page 67 of 100 Page ID #:708
2
3
4
5
6
7
8
9
10
11
12
Panel C: Severe Malnutrition Regression Results by Principal Diagnosis Bin
13
14
15
16
17
18
19
20
21
22
23
24
25 (ii) Aggregate Fixed Effect Regression Model
26 90. Second, Relator also ran a regression to calculate the cumulative effect
27 of Providence’s rate of Misstated MCCs across all claims in the relevant patterns.
28 The same regression described in Equation 1 is used for this analysis, except Relator
394151.1 64 Case No. 2:17-cv-01694-PSG-SS
SECOND AMENDED COMPLAINT
Case 2:17-cv-01694-PSG-SS Document 38 Filed 08/10/18 Page 68 of 100 Page ID #:709
1 runs one regression for all of the principal diagnosis codes in each MCC and adds a
2 control variable for the inpatient principal diagnosis category. This length of stay
3 variable is interacted with the principal diagnosis code to account for variation in the
4 expected length of stay given a principal diagnosis code.
5 91. As shown in Table 10, after controlling for other factors, the
6 Providence coefficient for the Misstated MCCs is 0.0808. This means that 8.08
7 percent of Providence claims are coded with one of the Misstated MCCs when they
8 would not have been coded as such at another hospital. Given the baseline usage
9 rate of the Misstated MCCs at other hospitals is 10.05 percent, Providence’s
10 calculated rate of Misstated MCCs is 18.13 percent. In other words, Providence’s
11 usage rate of the Misstated MCCs is 180.40% that of other hospitals, even after
12 controlling for patient, medical and demographic characteristics. This result is
13 statistically significant with more than 99.9999 percent confidence—i.e., almost
14 certainly not random.
15 92. Not surprisingly, the individual coefficients for encephalopathy,
16 respiratory failure, and severe malnutrition are also large in magnitude.
17 Providence’s usage rate for encephalopathy was 208.57% of the rate at other
18 hospitals, respiratory failure was 154.68%, and severe malnutrition was 223.92%.15
19
20
21
22
15
For robustness analysis, Relator also considered the possibility that certain
23
surgical procedure codes or admission sources (such as being admitted from the
24 emergency room) might explain the higher rates of Misstated MCCs at Providence.
Relator ran the fixed-effect regression analysis while also including controls for
25
surgical procedures and admission source. With these coefficients, Providence’s rate
26 of Misstated MCCs were 179.01% relative to other non-Providence hospitals.
Similarly, Providence’s usage rate for encephalopathy was 207.43% of the rate at
27
other hospitals, respiratory failure was 149.83%, and severe malnutrition was
28 217.80%.
394151.1 65 Case No. 2:17-cv-01694-PSG-SS
SECOND AMENDED COMPLAINT
Case 2:17-cv-01694-PSG-SS Document 38 Filed 08/10/18 Page 69 of 100 Page ID #:710
23
24 (iii) System and Hospital Residuals for Misstated
25 MCCs
26 93. Third, another regression method to analyze Providence’s coding of
27
28 16
LOS stands for length of stay.
394151.1 66 Case No. 2:17-cv-01694-PSG-SS
SECOND AMENDED COMPLAINT
Case 2:17-cv-01694-PSG-SS Document 38 Filed 08/10/18 Page 70 of 100 Page ID #:711
1 Figure 16. Average Unexplained Misstated MCC Rate for Each Hospital System and Individual
Hospital.
2 The following figure plots the results of the regression from Equation 1, except one regression was run for
all principal diagnosis bins, the control variable for principal diagnosis code was added to the regression,
3 and the Providence fixed effect variable was removed. All other variables included are the same. The graph
in Panel A is based on 729 hospital systems with at least 10,000 claims from 2011 through June 2017. The
4 graph in Panel B is based on 3240 hospitals with at least 500 claims during the same time period. The small
vertical lines off of the points represent the confidence interval for each system’s unexplained use of
5 Misstated MCCs.
6 Panel A: Average Residual of Any of the Misstated MCCs for Hospital Systems
7
8
9
10
11
12
13
14
15
16
17
Panel B: Average Residual of Any of the Misstated MCCs for Individual Hospitals
18
19
20
21
22
23
24
25
26
27
28
394151.1 68 Case No. 2:17-cv-01694-PSG-SS
SECOND AMENDED COMPLAINT
Case 2:17-cv-01694-PSG-SS Document 38 Filed 08/10/18 Page 72 of 100 Page ID #:713
1 95. Taken together, Relator’s regression analysis shows that the excessive
2 rates of Misstated MCCs was not due to unique patient demographic or health
3 characteristics but were specifically caused by the Defendants’ practices.
4 (b) Providence’s Merger with St. Joseph Health
5 Demonstrates that Providence Management Causes the
6 Excessive MCC Rates
7 96. Additionally, Providence’s responsibility for the fraudulent conduct can
8 be proven using causal methods, which are often used in economics, finance and
9 other applications to assess the extent to which an effect can be identified to be
10 caused, and not merely associated with, other explanatory variables.17 A common
11 causal econometric methodology is the use of discontinuity analysis, which can be
12 applied when there is a sudden change in the effect that one wishes to examine.18 A
13 discontinuity analysis is able to determine whether there is a statistically significant
14 sudden increase in the rate of a Misstated MCC due to an additional explanatory
15 variable, such as a change in ownership or management of a hospital. One common
16 causal econometric methodology that can be used to prove Providence’s fraudulent
17 conduct is known as a Regression of Discontinuity Design (RDD), which can be
18 applied to examine a sudden change in an effect in order to infer a causal
19 relationship.19 In this case, Providence’s causal influence on the rate of coding
20
17
21 “The notion of ceteris paribus—that is, holding all other (relevant) factors fixed—
is at the crux of establishing a causal relationship,” Jeffrey Wooldrige,
22
“Econometric Analysis of Cross Section and Panel Data” at 3, The MIT Press,
23 second edition, 2010.
24 18
See Angrist, J.D. and Pischke, J.S., “Mostly Harmless Econometrics: An
25 Empiricist’s Companion” at 251–253, Princeton University Press, 2009.
26 19
“CITS… produce[s] causally valid inferences about program impacts.” M.
Somers, P. Zhu, R. Jacob, H. Bloom, The Validity and Precision of the Comparative
27
Interrupted Time Series Design and the Difference-in-Difference Design in
28 Educational Evaluation, MDRC Working Paper on Research Methodology, 2013.
394151.1 69 Case No. 2:17-cv-01694-PSG-SS
SECOND AMENDED COMPLAINT
Case 2:17-cv-01694-PSG-SS Document 38 Filed 08/10/18 Page 73 of 100 Page ID #:714
1 Figure 17. Rate of Misstated MCC Coding by Month for Providence, St. Joseph, and Other Hospitals.
This figure shows the rate at which St. Joseph, Providence, and other hospitals are using one of the three
2 Misstated MCCs over time. St. Joseph is in red, Providence is in black, and other hospitals are in blue dots.
The dashed line on the left is the date at which the first contract was signed and the dashed line on the right
3 is when the agreement went into effect. St. Joseph’s rate of the three Misstated MCCs begins to increase when
the contract is signed with Providence and eventually has a similar usage rate after the agreement takes effect.
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
98. Relator conducted further analysis to show that the change in the rate of
19
Misstated MCCs was caused by the merger as opposed to other factors, such as
20
changing patient health and demographics, by using Regression of Discontinuity
21
Design (RDD). The red line in Figure 18 shows the results of the RDD analysis.
22
Here, Relator can quantify the jump that occurred immediately upon the merger.
23
Specifically, Relator calculates that the rate of Misstated MCCs jumped 3.92% as a
24
result of the merger. This calculation is based on the RDD formula shown in
25
Equation 2 below, which allowed Relator to compare the differences in coding
26
behavior at St. Joseph before and after the merger with Providence, relative to the
27
rate of Misstated MCCs nationwide. The advantage of this approach is that it
28
394151.1 71 Case No. 2:17-cv-01694-PSG-SS
SECOND AMENDED COMPLAINT
Case 2:17-cv-01694-PSG-SS Document 38 Filed 08/10/18 Page 75 of 100 Page ID #:716
1 allows Relator to identify and quantify not only the short-term effect of management
2 change on the immediate increase in coding misstated codes but also the long-term
3 effect in the post-contract trend of upcoding. This result was statistically significant
4 with a probability the jump is due to random chance being less than 1 in 100
5 million.
6
7 Figure 18. RDD of Misstated MCCs for Providence’s Merger with St. Joseph.
This figure shows the rate at which St. Joseph and other hospitals are using one of the three Misstated
8 MCCs over time. St. Joseph is in red and other hospitals are in blue dots. The dashed line on the left is the
date at which the first contract was signed and the dashed line on the right is when the agreement went into
9 effect. St. Joseph’s rate of the three Misstated MCCs begins to increase when the contract is signed with
Providence and is higher than the nationwide average after the agreement takes effect.
10
11
12
13
14
15
16
17
18
19
20
21
22
23 99. The increase in the rate of Misstated MCCs at St. Joseph upon merger
24 with Providence raises further questions. Namely, is it possible the increase in
25 Misstated MCCs is due to other less fraudulent factors? For example, is it possible
26 St. Joseph simply started taking on sicker patients that are more likely to have these
27 Misstated MCCs? Relator considered these possible explanations by re-running the
28 RDD described above, but by controlling for a variety of patient characteristics
394151.1 72 Case No. 2:17-cv-01694-PSG-SS
SECOND AMENDED COMPLAINT
Case 2:17-cv-01694-PSG-SS Document 38 Filed 08/10/18 Page 76 of 100 Page ID #:717
1 which might explain the jump in Misstated MCCs. Specifically, Relator controlled
2 for all of the characteristics discussed in Equation 1. This analysis allowed Relator
3 to isolate the marginal increase in the coding of Misstated MCCs that can be
4 attributed directly to the merger, beyond what might be explained by changes in the
5 underlying patient population. After controlling for potential alternative
6 explanations, Relator calculated that the merger led to a 3.99% increase in the rate
7 of Misstated MCCs coded on claims at St. Joseph. This result is highly significant
8 as the probability the increase is due to random chance is less than 1 in 100 million.
9 100. Relator’s calculation is based on the RDD formula shown in Equation 2
10 below, which allowed Relator to compare the differences in coding behavior at St.
11 Joseph before and after the merger with Providence, relative to the rate of Misstated
12 MCCs nationwide. The advantage of this approach is that it allows Relator to
13 identify and quantify not only the short-term effect of management change on the
14 immediate increase in coding misstated codes but also the long-term effect in the
15 post-contract trend of excessive rates of Misstated MCCs.
16
17 Equation 2. Relator’s Regression of Discontinuity Design Model.
The following equation presents the RDD model used by relator. The aggregate short-run and long-run
18 effect of management change on the level of upcoding at St. Joseph is estimated through the variable
. . This represents the jump in the rate of Misstated MCC upcoding while assuming that the coding
19 rate will continue to increase at the same rate as the nation-wide average (as represented by the slope of the
lines in Figure 18). The variable . represents the controls listed in Equation 1. When used without
20 controls, the outcome variable MCCit represents the monthly average MCC rate at St. Joseph in month t.
When run with controls, the outcome variable MCCit represents the binary value for existence or non-
21 existence of MCC in a specific claim i at St. Joseph in month t.
22 . . . . .
23
24 (c) Attending Physicians are not Responsible for the
25 Excessively High Rates of Misstated MCCs
26 101. Relator also considered whether the excessively high rates of Misstated
27 MCCs could be caused by the preferences or treatment decisions of physicians who
28 work with patients at Providence hospitals, as opposed to some system-wide
394151.1 73 Case No. 2:17-cv-01694-PSG-SS
SECOND AMENDED COMPLAINT
Case 2:17-cv-01694-PSG-SS Document 38 Filed 08/10/18 Page 77 of 100 Page ID #:718
1 103. Figure 20 shows that this tendency to have higher rates of Misstated
2 MCCs at Providence is not limited to a few doctors but is systemic. In the following
3 figure, doctors with the same rate of Misstated MCCs at Providence and other
4 hospitals would be clustered along the 45-degree line, whereas doctors with higher
5 rates of Misstated MCCs at Providence would be to the right of the 45-degree line.
6 As shown in Figure 20, 696 out of 1,010 doctors (or 68.9 percent) had a higher
7 coding rate of the Misstated MCCs at Providence than at other hospitals.
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
394151.1 75 Case No. 2:17-cv-01694-PSG-SS
SECOND AMENDED COMPLAINT
Case 2:17-cv-01694-PSG-SS Document 38 Filed 08/10/18 Page 79 of 100 Page ID #:720
1 Figure 20. Rate of Any of the Misstated MCCs for Common Doctors at Providence Versus Other
Hospitals.
2 The following figure compares the rate of Misstated MCCs for common doctors at Providence versus other
hospitals. In the graph, the red circles to the right of the 45-degree line represent doctors who have higher
3 upcoding at Providence and the blue circles represent doctors who have higher upcoding at other non-
Providence hospitals. Only doctors with at least 11 claims at Providence and 11 claims at a non-Providence
4 hospital are represented in this figure.
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
104. This result still holds when looking just at the individual Misstated
26
MCCs. For doctors that serve at both Providence and other hospitals, the rate of
27
encephalopathy at Providence was 13.68 percent, while the rate of encephalopathy
28
394151.1 76 Case No. 2:17-cv-01694-PSG-SS
SECOND AMENDED COMPLAINT
Case 2:17-cv-01694-PSG-SS Document 38 Filed 08/10/18 Page 80 of 100 Page ID #:721
1 Figure 22. Rate of Encephalopathy for Common Doctors at Providence Versus Other Hospitals.
The following figure compares the rate of encephalopathy for common doctors at Providence versus other
2 hospitals. In the graph, the red circles to the right of the 45-degree line represent doctors who have higher
upcoding of encephalopathy at Providence and the blue circles represent doctors who have higher upcoding
3 at other non-Providence hospitals. Only doctors with at least 11 claims at Providence and 11 claims at a
non-Providence hospital are represented in this figure.
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25 106. For doctors that serve at both hospitals, the rate of respiratory failure at
26 Providence was 30.92 percent, while the rate of respiratory failure at other hospitals
27 was 22.88 percent, as demonstrated in Figure 23 below. This suggests that a doctor
28 was 135.1% as likely to diagnosis a patient with respiratory failure when treating the
394151.1 78 Case No. 2:17-cv-01694-PSG-SS
SECOND AMENDED COMPLAINT
Case 2:17-cv-01694-PSG-SS Document 38 Filed 08/10/18 Page 82 of 100 Page ID #:723
1 patient at Providence than when the same doctor was treating a patient at a non-
2 Providence hospital.21
3
4 Figure 23. Rate of Respiratory Failure at Providence Relative to Other Hospitals for Claims with
Common Doctors.
5 The following figure includes any claims for doctors with at least 10 claims at Providence and 10 claims at
a non-Providence hospital from 2011 through June 2017. Even with doctors that work at both hospitals,
6 Providence had a respiratory failure rate of 30.92 percent, while those same doctors only use respiratory
failure on 22.88 percent of claims while at other hospitals. The analysis is based on 357 doctors with 10
7 claims at Providence and 10 claims at a non-Providence hospital. In total these doctors had more than
18,000 claims at Providence and more than 28,000 claims at other hospitals.
8
9
10
11
12
13
14
15
16
17
18
19
20 107. Figure 24 shows that a significant number of doctors had higher rates
21 of respiratory failure when they worked at Providence than at other hospitals.
22 Specifically, the figure shows that 232 doctors out of 357 doctors considered (or
23 65.0 percent) used a higher rate at Providence.
24
25
26
21
This general trend still holds when looking at any doctor that has at least one
27
claim at Providence and one claim at a non-Providence hospital. Specifically, the
28 rate of respiratory failure is 29.62% at Providence and 22.39% at other hospitals.
394151.1 79 Case No. 2:17-cv-01694-PSG-SS
SECOND AMENDED COMPLAINT
Case 2:17-cv-01694-PSG-SS Document 38 Filed 08/10/18 Page 83 of 100 Page ID #:724
1 Figure 24. Rate of Respiratory Failure for Common Doctors at Providence Versus Other Hospitals.
The following figure compares the rate of respiratory failure for common doctors at Providence versus other
2 hospitals. In the graph, the red circles to the right of the 45-degree line represent doctors who have higher
upcoding of respiratory failure at Providence and the blue circles represent doctors who have higher
3 upcoding at other non-Providence hospitals. Only doctors with at least 11 claims at Providence and 11
claims at a non-Providence hospital are represented in this figure.
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25 108. For doctors that serve at both hospitals, the rate of severe malnutrition
26 at Providence was 6.59 percent, while the rate of severe malnutrition at other
27 hospitals was 3.13 percent, as demonstrated in Figure 25 below. This indicates that
28 a doctor was 210.5% as likely to diagnosis a patient with severe malnutrition when
394151.1 80 Case No. 2:17-cv-01694-PSG-SS
SECOND AMENDED COMPLAINT
Case 2:17-cv-01694-PSG-SS Document 38 Filed 08/10/18 Page 84 of 100 Page ID #:725
1 treating the patient at Providence than when the same doctor was treating a patient at
2 a non-Providence hospital.22
3
4 Figure 25. Rate of Severe Malnutrition at Providence Relative to Other Hospitals for Claims with
Common Doctors.
5 The following figure includes any claims for doctors with at least 10 claims at Providence and 10 claims at
a non-Providence hospital from 2011 through June 2017. Even with doctors that work at both hospitals,
6 Providence had a severe malnutrition rate of 6.59 percent, while those same doctors only use severe
malnutrition on 3.13 percent of claims while at other hospitals. The analysis is based on 155 doctors with 10
7 claims at Providence and 10 claims at a non-Providence hospital. In total these doctors had 5,751 claims at
Providence and 7,708 claims at other hospitals.
8
9
10
11
12
13
14
15
16
17
18
19
20 109. Figure 26 shows that a significant number of doctors had higher rates
21 of severe malnutrition when they worked at Providence than at other hospitals. For
22 example, the Figure shows that 107 doctors out of 155 doctors considered (or 69.0
23 percent) used a higher rate at Providence.
24
25
26
22
This general trend still holds when looking at any doctor that has at least one
27
claim at Providence and one claim at a non-Providence hospital. Specifically, the
28 rate of severe malnutrition is 6.86% at Providence and 3.76% at other hospitals.
394151.1 81 Case No. 2:17-cv-01694-PSG-SS
SECOND AMENDED COMPLAINT
Case 2:17-cv-01694-PSG-SS Document 38 Filed 08/10/18 Page 85 of 100 Page ID #:726
1 Figure 26. Rate of Severe Malnutrition for Common Doctors at Providence Versus Other Hospitals.
The following figure compares the rate of severe malnutrition for common doctors at Providence versus
2 other hospitals. In the graph, the red circles to the right of the 45-degree line represent doctors who have
higher upcoding of severe malnutrition at Providence and the blue circles represent doctors who have higher
3 upcoding at other non-Providence hospitals. Only doctors with at least 11 claims at Providence and 11
claims at a non-Providence hospital are represented in this figure.
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25 110. Relator identified the specific doctors that had higher rates of
26 encephalopathy at Providence relative to other hospitals. Table 11 below lists the
27 ten doctors with the largest disparity in encephalopathy upcoding when they worked
28 at Providence compared to when they worked at other hospitals. Each of these
394151.1 82 Case No. 2:17-cv-01694-PSG-SS
SECOND AMENDED COMPLAINT
Case 2:17-cv-01694-PSG-SS Document 38 Filed 08/10/18 Page 86 of 100 Page ID #:727
1 a patient being diagnosed with a Misstated MCC, beyond what can be explained by
2 patient characteristics, demographic characteristics, as well as the individual doctor.
3 As shown in Table 12, Providence’s rate of any MCC, after these controls, was
4 135.87 percent of the rate at other hospitals, and Providence’s encephalopathy rate
5 was 138.86 percent of the rate at other hospitals.25
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
25
23 Relator also considered whether the behavior of these doctors is due to their
tendency to provide certain procedures at certain hospitals. To do this, Relator also
24 added control variables for the procedure codes and the admission status to identify
25 admissions as from the emergency room, elective, or urgent. For any Misstated
MCC and for Encephalopathy, the coefficients were 0.0479 and 0.0374 respectively.
26 In other words, Providence’s rate of Misstated MCC was 134.71 percent of other
27 hospitals, and rate of encephalopathy was 138.76 percent of other hospitals among
claims with common doctors.
28
394151.1 84 Case No. 2:17-cv-01694-PSG-SS
SECOND AMENDED COMPLAINT
Case 2:17-cv-01694-PSG-SS Document 38 Filed 08/10/18 Page 88 of 100 Page ID #:729
1 Table 12. Fixed Effect Regression Results After Controlling for Attending Physician.
Relator used a linear regression to analyze approximately 333,000 claims with common doctors at
2 Providence and other hospitals. The results are presented in the following table. The coefficient is listed
first, and the p-value is in parenthesis, which represents the statistical significance of the coefficient. A
3 lower p-value means the result is more statistically significant. Coefficients were not included for
categorical variables. Results were only included
4
Any Misstated MCC Encephalopathy
5
-0.0019 -0.0006
6 Poverty Rate
(<0.0001) (0.0554)
7 0.0035 0.0034
Unemployment Rate
(<0.0001) (<0.0001)
8 -0.0419 -0.0178
Log (Median Income)
(<0.0001) (0.0094)
9
0.0004 0.0005
No High School Diploma Rate
(0.0079) (0.0003)
10
0.4019 0.1163
Intercept
11 (<0.0001) (0.1612)
Principal Diagnosis Yes Yes
12 Principal Diagnosis X Log (LOS26) Yes Yes
Season Control Variables Yes Yes
13 Age Control Variables Yes Yes
Sex Control Variables Yes Yes
14 Race Control Variables Yes Yes
Discharge Status Group Controls Yes Yes
15 Principal Diagnosis Category Controls Yes Yes
RUCC Control Yes Yes
16 Doctor Control Variables Yes Yes
0.0495 0.0375
Providence Dummy Variable
17 (<0.0001) (<0.0001)
Nationwide Average 13.80% 9.65%
18
Providence Rate 18.75% 13.40%
19
Providence Relative Effect 135.87% 138.86%
20
21 112. This analysis shows that the fraudulent upcoding was not caused by
22 tendencies of certain doctors that treat patients at Providence but was instead caused
23 by coding practices that were implemented specifically at Providence with the
24 assistance of JATA.
25
26
27
28 26
LOS stands for length of stay.
394151.1 85 Case No. 2:17-cv-01694-PSG-SS
SECOND AMENDED COMPLAINT
Case 2:17-cv-01694-PSG-SS Document 38 Filed 08/10/18 Page 89 of 100 Page ID #:730
1 Figure 27. Rate of Any of The Misstated MCCs at Providence Relative to Other Hospitals for Claims
with Common Patients.
2 The following figure includes any claims for common patients between Providence and a non-Providence
hospital from 2011 through June 2017. Even with patients at both hospitals, Providence used one of the
3 Misstated MCCs on 23.83 percent of claims, while those same patients only have one of the Misstated
MCCs on 16.41 percent of claims while at other hospitals. The analysis is based on 850 patients with 5
4 claims at Providence and 5 claims at a non-Providence hospital. In total these patients had 7,078 claims at
Providence and 8,536 claims at other hospitals.
5
6
7
8
9
10
11
12
13
14
15
16
17 115. For patients that were treated at both hospitals, the rate of
18 encephalopathy at Providence was 19.81 percent, while the rate of encephalopathy
19 at other hospitals was 13.79 percent, as demonstrated in Figure 28 below. This
20 suggests that a patient was 143.7% as likely to be diagnosed with encephalopathy
21 when being treated at Providence than when the same patient was treated at a non-
27, 28
22 Providence hospital.
23
24
27
25 This general trend still holds when looking at any patient that has at least one
claim at Providence and one claim at a non-Providence hospital. Specifically, the
26 rate of encephalopathy is 34.95% at Providence and 22.87% at other hospitals.
27 28
These results were not presented for respiratory failure and severe malnutrition
28 due to the small number of common patients available for analysis.
394151.1 87 Case No. 2:17-cv-01694-PSG-SS
SECOND AMENDED COMPLAINT
Case 2:17-cv-01694-PSG-SS Document 38 Filed 08/10/18 Page 91 of 100 Page ID #:732
1 Figure 28. Rate of Encephalopathy at Providence Relative to Other Hospitals for Claims with
Common Patients.
2 The following figure includes any claims for patients with at least 5 claims at Providence and 5 claims at a
non-Providence hospital from 2011 through June 2017. Even with patients at both hospitals, Providence had
3 an encephalopathy rate of 19.81 percent, while those same patients only have encephalopathy on 13.79
percent of claims while at other hospitals. The analysis is based on 671 patients with 5 claims at Providence
4 and 5 claims at a non-Providence hospital. In total these patients had 5,584 claims at Providence and 6,615
claims at other hospitals.
5
6
7
8
9
10
11
12
13
14
15
16 116. As this analysis shows, even when looking at the same patient,
17 Providence has significantly higher rates of Misstated MCCs than other hospitals.
18 This shows that the upcoding behavior cannot be attributable to patient differences.
19 (e) Regional Factors do not Explain Why Providence Has
20 Higher Rates of MCCs
21 117. Relator also considered whether the high rates of MCC upcoding at
22 Providence hospitals might be due to the region in which Providence’s hospitals are
23 located. Although Relator has already controlled for a variety of county
24 demographic factors through the regression, Relator now compares the rate of
25 Misstated MCCs between Providence and other hospitals within each MSA.29
26 118. As shown in Table 13, Providence had a significantly higher rate of
27
28 29
Relator only included MSAs in which there were at least 2 comparison hospitals.
394151.1 88 Case No. 2:17-cv-01694-PSG-SS
SECOND AMENDED COMPLAINT
Case 2:17-cv-01694-PSG-SS Document 38 Filed 08/10/18 Page 92 of 100 Page ID #:733
1 Misstated MCCs in each MSA. As shown in Table 14, Providence had a higher rate
2 of encephalopathy than other hospitals in each MSA, with some having
3 encephalopathy rates that are nearly twice as high as the rate at other hospitals in the
4 same MSA.
5 119. Table 15 shows the rate of respiratory failure was higher for Providence
6 hospitals in all but one MSA. Table 16 shows that Providence had a higher rate of
7 severe malnutrition in all but one MSA, with Providence hospitals in the Spokane-
8 Spokane Valley MSA coding severe malnutrition at more than four times the rate as
9 other hospitals in the MSA.
10
11 Table 13. Rate of Misstated MCCs at Providence Versus Other Hospitals in the Same MSA.
This table compares the rate of Misstated MCCs in the suspicious patterns for the hospitals in Providence
12 and other hospitals within the same geographic region.
Providence Rate
13 Providence Nationwide Relative to Other
MSA MCC Rate MCC Rate Hospitals Probability
14 Los Angeles-Long Beach-Anaheim, CA 17.11% 12.68% 135% <0.0001
Medford, OR 16.04% 12.18% 132% <0.0001
15 Portland-Vancouver-Hillsboro, OR-WA 17.07% 13.21% 129% <0.0001
Seattle-Tacoma-Bellevue, WA 16.59% 11.84% 140% <0.0001
16 Spokane-Spokane Valley, WA 18.63% 10.91% 171% <0.0001
17 Table 14. Rate of Encephalopathy at Providence Versus Other Hospitals in the Same MSA.
This table compares the encephalopathy rate of the suspicious patterns for the hospitals in Providence and
18 other hospitals within the same geographic region.
19 Providence Nationwide Providence Rate
Encephalopat Encephalopathy Relative to Other
20 MSA hy Rate Rate Hospitals Probability
Los Angeles-Long Beach-Anaheim,
12.38% 9.56% 130% <0.0001
21 CA
Medford, OR 11.65% 9.32% 125% <0.0001
Portland-Vancouver-Hillsboro, OR-
22 WA
13.12% 8.77% 150% <0.0001
Seattle-Tacoma-Bellevue, WA 13.03% 7.73% 169% <0.0001
23 Spokane-Spokane Valley, WA 13.64% 6.99% 195% <0.0001
24 Table 15. Rate of Respiratory Failure at Providence Versus Other Hospitals in the Same MSA.
25 This table compares the respiratory failure rate of the suspicious patterns for the Providence hospitals and
other hospitals within the same geographic region.
26 Providence Nationwide Providence Rate
Respiratory Respiratory Failure Relative to
27 MSA Failure Rate Rate Other Hospitals Probability
Los Angeles-Long Beach-Anaheim, CA 26.53% 17.06% 156% <0.0001
28 Medford, OR 23.82% 16.10% 148% <0.0001
1 3. Economic Damages
2 122. Relator employed a robust and conservative methodology to quantify
3 the economic damages caused by the Defendants’ fraudulent coding
4 Encephalopathy, Respiratory Failure, and Severe Malnutrition. Relator has limited
5 this complaint to only the most extreme cases—i.e., where Providence used a
6 Misstated MCC code at two times the rate of comparable hospitals or at least three
7 percentage points of its entire patient population higher than other hospitals.
8 Additionally, only principal diagnosis bins where the excessive MCC usage rate was
9 statistically significant at a 99.9% rate—or almost certainly not random—were
10 considered fraudulent. The following describes Relator’s methodology for
11 aggregating the total dollar value of the fraud committed by Providence.
12 123. Relator employs a principal diagnosis bin-based regression
13 methodology for calculating damages. For each principal diagnosis bin, Relator re-
14 ran its fixed effect linear regression model discussed in Equation 1 but changed the
15 dependent variable to represent the additional revenue due to upcoding. For each
16 claim, Relator calculated the difference in the DRG weight between claims with
17 Misstated MCCs and claims without Misstated MCCs.30 Relator then multiplied
18 this difference in weights by the average base rate from 2011 through 2017, which
19 was $6,417.57.31 Within the regression for each principal diagnosis bin, the fixed
20
21
30
For claims that could have been a complication or without complication, Relator
22 took a weighted average of DRG weights for the two DRGs and weighted by
23 Providence’s historical distribution of severity levels. Approximately 16.3% of
claims were without complication and 83.7% of claims were with complication. If
24 Providence also upcodes using CC secondary diagnoses, the damage calculation
25 would be even more conservative.
26 31
The labor portion of the base rate was further adjusted by the average wage index
among Providence hospitals from 2011 through 2017, and the capital portion was
27
adjusted by the geographic adjustment factor over the same time period, to get a
28 more accurate calculation of additional revenue.
394151.1 91 Case No. 2:17-cv-01694-PSG-SS
SECOND AMENDED COMPLAINT
Case 2:17-cv-01694-PSG-SS Document 38 Filed 08/10/18 Page 95 of 100 Page ID #:736
1 effect for Providence represents the additional revenue Providence receives for the
2 Misstated MCCs after controlling for possible differences in patient, regional, and
3 claim characteristics. Relator further only attributed damages for the regression
4 results that were statistically significant at a 99.9% level, meaning there is a less in 1
5 in 1,000 chance the additional revenue received is due to random chance.
6 124. Based on this bin-based regression, Relator’s analysis shows that
7 Providence received an additional $188.1 million in false claims across all principal
8 diagnosis categories due to fraudulent MCC upcoding. Table 17 demonstrates the
9 additional revenue Providence received for false claims across each of the Misstated
10 MCCs.
11
12 Table 17. Damages by Specific Misstated MCCs.
14 Encephalopathy $153,556,713
Respiratory Failure $29,012,041
15 Severe Malnutrition $5,519,441
16 Total $188,088,196
1 relatively normal and would thus lead to a lower damage estimate.32 Indeed it is
2 overly conservative to compare Providence’s fraudulent behavior to other hospitals
3 also engaging in the same fraudulent activity; therefore Relator also undertook a
4 different methodology to identify hospital systems based on the amount of
5 fraudulent activity they have among all of their claims. If Relator were to remove
6 from comparison the top third of hospital systems identified to have excessively
7 billed Medicare and re-run the bin-based regression analysis, damages would total
8 $237.5 million.
9 127. Relator’s consideration of other possible explanations, such as claim
10 characteristics, patient characteristics, and doctor practices, demonstrates that the
11 excessive coding of Misstated MCCs is due to system-wide practices in place at
12 Providence, with the assistance of JATA. Additionally, the extremely high levels of
13 statistical significance of the analyses across a variety of comparative settings
14 indicate a nearly impossible probability that the practices are due to random chance.
15 Relator’s damages estimate of $188.1 million due to Providence’s fraudulent
16 upcoding is conservative and the estimate is robust when controlling for a variety of
17 factors.
18 128. Through its proprietary research and analysis, Relator also uncovered
19 that JATA’s encouragement and enabling of hospitals to code at excessive rates of
20 the Misstated MCCs extends beyond Providence hospitals. While the full extent of
21 JATA’s fraudulent activity is likely much larger, Relator estimates that JATA
22 assisted other hospitals to submit falsely inflated claims to Medicare of at least
23 $337.0 million.
24
25
26
32
As an example, the comparison set of hospitals includes Prime Healthcare
27
Services, Inc., which is currently being sued by the US Department of Justice under
28 the False Claims Act. See https://goo.gl/rZcES9.
394151.1 93 Case No. 2:17-cv-01694-PSG-SS
SECOND AMENDED COMPLAINT
Case 2:17-cv-01694-PSG-SS Document 38 Filed 08/10/18 Page 97 of 100 Page ID #:738
1 CAUSES OF ACTION
2 COUNT ONE
3 Violation of the False Claims Act arising from
4 falsifying patient diagnoses, complications, and comorbidities
5 (Against All Defendants)
6 129. Relator repeats and realleges each and every allegation contained above
7 as if fully set forth herein.
8 130. As described above, Defendants have submitted and/or caused to be
9 submitted false or fraudulent claims to Medicare by falsifying material information
10 concerning patient diagnoses, complications, and comorbidities; and by failing to
11 report and return overpayments from Medicare.
12 131. Defendants, by the conduct set forth herein, have violated:
13 a. 31 U.S.C. § 3729(a)(1)(A) by knowingly presenting, or causing to
14 be presented, false or fraudulent claims for payment or approval;
15 and/or
16 b. 31 U.S.C. § 3729(a)(1)(B) by knowingly making, using or causing
17 to be made or used, a false record or statement material to a false
18 or fraudulent claim; and/or
19 c. 31 U.S.C. § 3729(a)(1)(G) by knowingly making, using, or
20 causing to be made or used, a false record or statement material to
21 an obligation to pay or transit money or property to the
22 government, or knowingly concealing or knowingly and
23 improperly avoiding or decreasing an obligation to pay or transmit
24 money or property to the government.
25 132. Through use of Defendant JATA’s services as described above, the
26 Providence Defendants also conspired with JATA to defraud the federal
27 government, in violation of 31 U.S.C. § 3729(a)(1)(C), by knowingly and
28 systematically falsifying claims allowed or paid by the government.
394151.1 94 Case No. 2:17-cv-01694-PSG-SS
SECOND AMENDED COMPLAINT
Case 2:17-cv-01694-PSG-SS Document 38 Filed 08/10/18 Page 98 of 100 Page ID #:739
1 133. The United States has suffered and continues to suffer damages as a
2 direct proximate result of Defendants’ false or fraudulent claims.
3 COUNT TWO
4 Violations of the Federal False Claims Act arising from the Defendants’
5 violations of the Anti-Kickback Statute
6 (Against All Defendants)
7 134. Relator repeats and realleges each and every allegation contained above
8 as if fully set forth herein.
9 135. By offering JATA financial incentives for upcoding and making all or
10 some of JATA’s compensation contingent on increases in Providence’s Medicare
11 revenue, Providence provided JATA with incentives to arrange for or recommend
12 services for which payment may be made in whole or in part under Medicare, in
13 violation of the 42 U.S.C. § 1320a-7b(b) (the “Anti-Kickback Statute”).
14 136. Defendant JATA’s consulting engagement with Providence is not
15 protected under the existing “safe harbor” provisions of the Anti-Kickback Statute.
16 137. As a material prerequisite to any recovery or reimbursement from
17 Medicare, the Providence Defendants expressly or impliedly certified their
18 compliance with the Anti-Kickback Statute.
19 138. The Providence Defendants violated the False Claims Act by
20 submitting Medicare claims knowing or recklessly disregarding that the Providence
21 Defendants were ineligible for reimbursement on such claims because of their
22 violations of the Anti-Kickback Statute.
23 139. Defendant JATA violated the False Claims Act by causing Medicare
24 claims to be submitted by the Providence Defendants even though JATA knew or
25 recklessly disregarded that the Providence Defendants were ineligible for the
26 payments demanded because of their violations of the Anti-Kickback Statute.
27 140. Each Medicare claim submitted by the Providence Defendants that was
28 arranged or recommended by JATA is false and material because it is tainted by the
394151.1 95 Case No. 2:17-cv-01694-PSG-SS
SECOND AMENDED COMPLAINT
Case 2:17-cv-01694-PSG-SS Document 38 Filed 08/10/18 Page 99 of 100 Page ID #:740