Disruption of Olanzapine Therapy and Changes in Health Care Resource Utilization

advertisement

Disruption of Olanzapine Therapy and

Changes in Health Care Resource Utilization

Following a Medicaid Policy Change

James Signorovitch, Ph.D.

1

Ben-Hamadi, M.Sc.

1

; Howard Birnbaum, Ph.D.

1 ; Rym

Lawson, M.A.

2

; Andrew P. Yu, Ph.D.

; Daniel Ball, DrPH 2

1 ; Yohanne Kidolezi,

B.A.

1 ; David Kelley, B.A.

1 ; Glenn Phillips, Ph.D.

2 ; Anthony

1. Analysis Group, Inc., Boston, MA, USA

2. Eli Lilly and Company, Indianapolis, IN, USA

Study funding provided by Eli Lilly and Company

AcademyHealth Annual Research Meeting, June 30 th , 2009, Chicago

Background

Atypical antipsychotics (AAs) are frequent targets of cost containment measures

 AAs account for over 13% of Medicaid prescription drug spending

 About 40% of state Medicaid programs require prior authorization (PA) for AAs 1

Formulary policies limiting access to AAs have been associated with 2,3,4 :

 Reduced treatment compliance

 Increased use of medical services

 Increased use of other medications

2 | AcademyHealth Annual Research Meeting, June 30 th , 2009, Chicago

Background (cont.)

On July 11th, 2005, Florida Medicaid re-classified olanzapine as non-preferred

 Existing olanzapine users were not “grandfathered” – physicians had 60 days to transition olanzapine patients to a different antipsychotic

 The policy change was communicated via a public awareness campaign to physicians, patients, and pharmacists

 The policy change was rescinded just prior to the end of the

60-day transition period and was thus never fully implemented

3 | AcademyHealth Annual Research Meeting, June 30 th , 2009, Chicago

Objective

Assess the impact of the Florida Medicaid policy change among olanzapine-treated patients diagnosed with schizophrenia or bipolar disorder in terms of:

 Continuity of olanzapine therapy

 Health care resource use

 Health care reimbursements

4 | AcademyHealth Annual Research Meeting, June 30 th , 2009, Chicago

Data

De-identified Medicaid claims data from 01/2004-01/2006 in two states:

 Florida (FL):

 3.9 million lives

 Policy change affecting atypical antipsychotics on 7/11/2005

 New Jersey (NJ):

 2.7 million lives

 No policy change affecting atypical antipsychotics during the study period

5 | AcademyHealth Annual Research Meeting, June 30 th , 2009, Chicago

Data (cont.)

The data from both FL and NJ includes

 complete medical and pharmaceutical claims from both feefor-service and managed care organizations

 dates of services

 dates of prescription fills

 prescription days of supply

 Medicaid reimbursed amounts

All reimbursements were inflated to 2005 U.S. dollars using the medical component of the Consumer Price Index

6 | AcademyHealth Annual Research Meeting, June 30 th , 2009, Chicago

Methods: Outcome Measures

Disruption of olanzapine: prescription gap

≥15 days

 Switching: filling 2 or more prescriptions for another antipsychotic before or during the first olanzapine disruption

 Discontinuation: patients with disrupted olanzapine treatment who were not classified as switchers (i.e., those who did not establish use of other antipsychotics during the follow-up period after the olanzapine gap)

Use of other antipsychotics

7 | AcademyHealth Annual Research Meeting, June 30 th , 2009, Chicago

Methods: Outcome Measures (cont.)

Health care resource use:

 Hospitalization

 Emergency room (ER) visits

Health care reimbursements:

 Medical services

 Prescription drugs

 Total (medical services + drugs)

8 | AcademyHealth Annual Research Meeting, June 30 th , 2009, Chicago

Methods: Selection of Olanzapine Users in FL

Policy change index date

(7/11/2005)

6-month baseline period

 Continuous Medicaid eligibility

 ≥ 2 continuous prescriptions for olanzapine with the last one covering the index date

 ≥ 2 outpatient or ≥ 1 inpatient diagnosis for schizophrenia and / or bipolar disorder

 Aged 18-64 years on the index date

9 | AcademyHealth Annual Research Meeting, June 30 th , 2009, Chicago

6-month outcome period

 Continuous Medicaid eligibility

Methods: Pre-Policy Controls

FL

2005

Policy change

(7/11/2005)

2004

7/11/2004

10 | AcademyHealth Annual Research Meeting, June 30 th , 2009, Chicago

Methods: Reference State

FL

2005

Policy change

(7/11/2005)

2004

7/11/2004

11 | AcademyHealth Annual Research Meeting, June 30 th , 2009, Chicago

NJ

7/11/2005

7/11/2004

Methods: Matching within States

FL

2005

NJ

1:1 matching 1:1 matching

2004

Patients were matched in terms of their propensity to be in the 2005 vs. 2004 cohorts

 The propensity score was developed from baseline characteristics including demographics, comorbidities, prescription drug use, and health care service use

12 | AcademyHealth Annual Research Meeting, June 30 th , 2009, Chicago

Methods: Within-State Comparison of Outcomes

FL NJ

2005

1:1 matching

Comparison of outcomes

1:1 matching

Comparison of outcomes

2004

 Outcomes were compared between matched pairs within states using Wilcoxon signed rank tests for continuous variables and

McNemar’s test for dichotomous variables

13 | AcademyHealth Annual Research Meeting, June 30 th , 2009, Chicago

Methods: Between-State Comparison

FL

2005

NJ

1:1 matching

Comparison of outcomes

1:1 matching

Comparison of outcomes

2004

 Within-state changes in olanzapine disruption and health care resource use 2004-2005 were compared in FL vs. NJ

14 | AcademyHealth Annual Research Meeting, June 30 th , 2009, Chicago

Results: Baseline Characteristics in FL

Baseline

Characteristics

Age (years), mean ± s.d.

Before Matching

2004 2005

(N=7,257)

49.47 ± 13.72

(N=5,164)

49.91 ± 13.77

After Matching

2004

(N=4,255)

49.45 ± 13.50

Male, % (N)

Schizophrenic disorders, % (N)

Charlson

Comorbidity Index

(CCI), mean ± s.d.

53.6%

(3,890)

59.7%

(4,332)

0.63 ± 1.33

*P<0.05; **P<0.01; ***P<0.001

56.1%

(2,896)*

57.3%

(2,957)*

0.49 ± 1.16***

54.9%

(2,338)

57.3%

(2,436)

0.41 ± 0.94

2005

(N=4,255)

49.87 ± 13.50

55% (2,341)

57.6%

(2,452)

0.43 ± 1.01

15 | AcademyHealth Annual Research Meeting, June 30 th , 2009, Chicago

Results: Baseline Characteristics in FL (cont.)

Baseline Characteristics

Before Matching

2004 2005

After Matching

2004 2005

Medication Use, % (N)

Typical antipsychotics 11.5% (835)

Atypical antipsychotics 27% (1,959)

Anticholergenics 19.1% (1,384)

11.4% (586)

25.6% (1,322)

19.4% (1,000)

Benzodiazepines/ anxiolytics

58.1% (4,216) 51% (2,634)***

11% (469)

25% (1,065)

19% (807)

52.9% (2,250)

11.2% (477)

26.2% (1,114)

19.3% (820)

51.4% (2,187)

Antidepressants

Mood stabilizers

61.4% (4,452) 58.3% (3,010) 59.7% (2,540) 59.6% (2,534)

35.1% (2,548) 36.2% (1,869) 33.7% (1,433) 33.9% (1,443)

Medical Resource Use, % (N)

Hospitalization 16.5% (1,194) 13.9% (717)***

Emergency visits 21.2% (1,540) 18.2% (940)***

*P<0.05; **P<0.01; ***P<0.001

10.5% (447)

15.8% (672)

10.3% (437)

15.2% (648)

16 | AcademyHealth Annual Research Meeting, June 30 th , 2009, Chicago

Results: Disruption of Olanzapine Treatment

2004 2005

90%

80%

70%

60%

50%

40%

30%

20%

10%

0%

Cumulative Disruption Rates

Florida p < 0.001

0 1 2 3 4 5 6

Months Following Index Date

90%

80%

70%

60%

50%

40%

30%

20%

10%

0%

Cumulative Disruption Rates

New Jersey p = 0.23

0 1 2 3 4 5 6

Months Following Index Date

17 | AcademyHealth Annual Research Meeting, June 30 th , 2009, Chicago

Results: Switching from Olanzapine Treatment

2004 2005

45%

40%

35%

30%

25%

20%

15%

10%

5%

0%

Cumulative Switching Rates

Florida p < 0.001

0 1 2 3 4 5 6

Months Following Index Date

45%

40%

35%

30%

25%

20%

15%

10%

5%

0%

Cumulative Switching Rates

New Jersey p = 0.14

0 1 2 3 4 5 6

Months Following Index Date

18 | AcademyHealth Annual Research Meeting, June 30 th , 2009, Chicago

Results: Medication Use

Other Medication Use

Patients with Outcome, % (N)

2004 2005

Florida Medicaid (N=4,255)

Typical antipsychotics

Atypical antipsychotics 1

10.6% (450) 13.7% (581)

33.6% (1,428) 67.7% (2,882)

Change from

2004 to 2005

29.1%

101.8%

p-value

<.0001

<.0001

New Jersey Medicaid (N=2,680)

Typical antipsychotics 13.4% (358)

Atypical antipsychotics 1 27.9% (748)

14.0% (374)

26.4% (708)

4.5%

-5.3%

1.

aripripazole, clozapine, quetiapine, risperidone, ziprasidone (excludes olanzapine)

0.5143

0.2140

19 | AcademyHealth Annual Research Meeting, June 30 th , 2009, Chicago

Results: Hospitalization and ER Visits

Hospitalizations ER Visits

40%

40%

P = 0.95

30%

30%

P <0.001

20%

20%

P = 0.78

P < 0.001

10%

10%

0%

0%

2004 2005

Matched FL Cohorts

2004 2005

Matched NJ Cohorts

 Hospitalization and ER visits increased by 20% following the policy change in FL Medicaid

 No significant changes in these outcomes in NJ Medicaid during the same time periods

20 | AcademyHealth Annual Research Meeting, June 30 th , 2009, Chicago

Results: Per-Patient Reimbursements in FL

Per-Patient Reimbursements

Medical services 2

Hospitalization

Emergency visits

Prescription drugs

Olanzapine

Other atypicals

Total

U.S. Dollars (2005), Mean ± SD

2004

(N=4,255)

4190 ± 7747

709 ± 3634

33 ± 128

4554 ± 2616

2210 ± 1379

494 ± 958

8744 ± 8213

2005

(N=4,255)

4471 ± 9206

966 ± 5499

40 ± 133

4255 ± 2727

1316 ± 1304

1046 ± 1203

8726 ± 9673

Change from

2004 to 2005 P-value 1

281

256

6

-299

-895

552

-18

0.1026

0.0468

0.0926

<.0001

<.0001

<.0001

0.1374

1. Dichotomous outcomes were compared using McNemar's test; continuous outcomes were compared using the Wilcoxon signed rank test.

2. Includes reimbursements associated with all Medicaid claims including home and community health services.

21 | AcademyHealth Annual Research Meeting, June 30 th , 2009, Chicago

Conclusions

For olanzapine-treated patients diagnosed with schizophrenia or bipolar disorder, removal of olanzapine from Florida Medicaid’s preferred drug list without “grandfathering” existing users was associated with:

 Increased disruption of olanzapine treatment

 Increased use of other atypical antipsychotics

 Increased rates of hospitalization and ER visits

 Decreased reimbursements for prescription drugs largely offset by increased reimbursements for hospitalization

 No significant change in total medical costs

22 | AcademyHealth Annual Research Meeting, June 30 th , 2009, Chicago

Conclusions (cont.)

Findings support recommendations of the National Association of State Mental Health Program Directors 5

 “Patients should not be forced to switch medications due to changes in formulary policy”

 “‘Grandfathering’ is the recommended practice for individuals stabilized on a nonformulary antipsychotic medication to minimize risk of relapse and support continuity of care”

23 | AcademyHealth Annual Research Meeting, June 30 th , 2009, Chicago

Thank You

| AcademyHealth Annual Research Meeting, June 30 th , 2009, Chicago

References

1.

Polinski, J. M., P. S. Wang, et al. (2007). "Medicaid's Prior Authorization

Program and Access to Atypical Antipsychotic Medications." Health Affairs

26(3): 750-760.

2.

K.T. Mueser and S.R. McGurk, "Schizophrenia," Lancet 363, no. 9426

(2004): 2063–2072

3.

Zeber, J., K. L. Grazier, et al. (2007). "The Effect of a Medication

Copayment Increase in Veterans with Schizophrenia." The American

Journal of Managed Care 13(6 (Part 2)): 335-346

4.

Soumerai, S. B., T. J. McLaughlin, et al. (1994). "Effects of limiting Medicaid drug-reimbursement benefits on the use of psychotropic agents and acute mental health services by patients with schizophrenia." New England

Journal of Medicine 331(10): 650-5.

5.

Parks, J., A. Radke, et al. (2008). "Principles of Antipsychotic Prescribing for Policy Makers, Circa 2008. Translating Knowledge to Promote

Individualized Treatment." Schizophr Bull: sbn019.

25 | AcademyHealth Annual Research Meeting, June 30 th , 2009, Chicago

Download