Post-market Authorized Generic Evaluation (PAGE) · PDF filePost-market Authorized Generic Evaluation (PAGE) U01FD005272-02 November 18th, 2016 ... Health Services Use & Outcomes 9

Post-market Authorized Generic Evaluation (PAGE)

U01FD005272-02 November 18th, 2016

Peggy L. Peissig

Richard Berg

Michael Caldwell

James Linneman

Richard Hansen

Jingjing Qian

Motiur Rahman

Ning Cheng

Yasser Alatawi

David Page

Enrique Seoane-Vazquez

Project Overview

2

Specific Aim 1

• To determine and compare switchback rates, medical service utilization, and clinical outcomes between authorized generics (AGs) and generics using healthcare claim data with electronic medical records.

3

Aim 1 Methods

• Study design: a retrospective cohort study

• Study population: Marshfield Clinic (MC) Security Health Plan (SHP) claims data with linked electronic health records (EHR) data in 1999-2014

• Inclusion criteria:– Continuous enrollment (CE) in 6-mon prior to generic introduction

through at least the first Rx fill after generic availability

– At least 1 brand Rx in 6-mon prior and 1 Rx fill of a medication in the therapeutic area within 12-mon after generic availability

– At least 1 MC healthcare encounter/year during eligible periods

4

Aim 1 Methods (cont’d)

Branded treatment initiation

Generic entry

Generic switch

Non-switchers: stay on brand

SwitchbackIndex date for generic

switch

Index date for switchback

Switchers: stay on generic

Switchers: switchback to brand

Pre-index

Brand

Brand

Generic

Generic switch: switch from a brand drug to an authorized or independent generic drug within 30 months following generic entry

Switchback: among those who had a brand to generic switch, generic to brand switchback rates were calculated in 30 months following the index switch date

5

Aim 1 Methods (cont’d)

• Revised drugs of interest:

Drugs Generic Switch Switchback

alendronate x x

amlodipine x x

citalopram x x

gabapentin x x

glimepiride x

losartan x

metformin x

paroxetine x x

sertraline x x

simvastatin x x

6

Rates of Switching

Drugs Switch Type

Brand to AG Brand to OG

All drugs (n=5234) 1138 (23%) 3762 (77%)

Alendronate (n=930) 41 (5%) 832 (95%)

Amlodipine (n=1487) 289 (20%) 1156 (80%)

Citalopram (n=813) 74 (10%) 670 (90%)

Gabapentin (n=279) 25 (11%) 199 (89%)

Paroxetine (n=669) 302 (48%) 328 (52%)

Sertraline (n=730) 278 (40%) 417 (60%)

Simvastatin (n=636) 176 (30%) 408 (70%)

7

*Factors associated with a higher likelihood of generic switch in the overall drug cohort included pre-index defined daily dose and all-cause hospitalization

Time from generic entry to generic switch by drug

Rates of Switching Back to Brand

Factors Brand to Generic Switchback

Hazard

Ratio

95% Confidence

Interval

IG (REF=AG) 0.86 0.65-1.15

Age1.02* 1.01-1.02

Male 0.59* 0.44-0.80

Proportion of pre-index brand

medication use 2.43* 1.45-4.07

Defined daily dose prior to switching 0.85 0.72-1.01

Charlson comorbidity index 1.02 0.90-1.16

Pre-index hospitalization0.87 0.52-1.48

Pre-index ED visit 1.02 0.67-1.56

Count of pre-index outpatient visits 1.02 1.00-1.02

8

*P<0.05, multivariable Cox proportional hazards model

Brand switchback rate from AG vs. OG

Health Services Use & Outcomes

9

*Medical service utilization and outcomes were measured during a 12-month period after generic switch. Generalized logistic regression was used to model binary outcomes and negative binomial regression were used to mode count outcomes. Age, defined daily dose, and Charlson score were controlled as covariates in these models.

Adjusted medical service utilization and outcomes for AG vs. OG users

Outcome Estimate

Lower

CI

Upper

CI P-Value

1.05 1 1.1 0.071

1.08 0.9 1.29 0.395

All-cause emergency department visits

Any visit 1.33 1.11 1.61 0.003

Number per year 1.23 1.02 1.47 0.026

All-cause hospitalizations

Any visit 1.14 0.91 1.43 0.257

Number per year 1.09 0.81 1.46 0.582

Medication discontinuation 0.95 0.8 1.12 0.508

Number of all-cause

outpatient visists per year

Number of all-cause urgent

care visits per year

1Estimate

0.5 2Favors OG Favors AG

Specific Aim 2

• To analyze brand versus generic adverse event reporting rates in the FDA Adverse Event Reporting System (FAERS)

10

Methods

• Data– FAERS data 2004-2014

• Approach– Method 1 – verbatim assignment by name

– Method 2 – assign reports to manufacturer• Exclude direct reports

• Sensitivity analyses around which reports to include, such as primary suspect, serious, US only, validated

• Analyses– Disproportionality analyses

– Segmented regression11

Drugs

• Alendronate

• Amlodipine

• Azithromycin

• Carbamazepine XR

• Escitalopram

• Lamotrigine

• Leflunomide

• Losartan

• Metoprolol XR

• Modafinil

• Oxcarbazepine

• Sertraline

• Venlafaxine ER

• Zolpidem

12

Events

• Coded by Preferred Terms (PT) or Standardized MedDRA Query (SMQ)

• Drug specific events defined by label

• Universal events

13

- Accidents & injuries - Lack of efficacy

- Anaphylactic reactions - Ischemic heart disease

- Embolic & thrombotic events - Hematopoietic cytopenias

- Hemorrhages - Death

Report Type as Percentage of Total

14

0

20

40

60

80

100

120

US

US,

PS

US,

Ser

iou

s, v

alid

, PS US

US,

PS

US,

Ser

iou

s, v

alid

, PS US

US,

PS

US,

Ser

iou

s, v

alid

, PS US

US,

PS

US,

Ser

iou

s, v

alid

, PS US

US,

PS

US,

Ser

iou

s, v

alid

, PS US

US,

PS

US,

Ser

iou

s, v

alid

, PS US

US,

PS

US,

Ser

iou

s, v

alid

, PS US

US,

PS

US,

Ser

iou

s, v

alid

, PS US

US,

PS

US,

Ser

iou

s, v

alid

, PS US

US,

PS

US,

Ser

iou

s, v

alid

, PS US

US,

PS

US,

Ser

iou

s, v

alid

, PS US

US,

PS

US,

Ser

iou

s, v

alid

, PS US

US,

PS

US,

Ser

iou

s, v

alid

, PS US

US,

PS

US,

Ser

iou

s, v

alid

, PS US

US,

PS

US,

Ser

iou

s, v

alid

, PS US

US,

PS

US,

Ser

iou

s, v

alid

, PS US

US,

PS

US,

Ser

iou

s, v

alid

, PS US

US,

PS

US,

Ser

iou

s, v

alid

, PS US

US,

PS

US,

Ser

iou

s, v

alid

, PS

-8 -7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7 8 9 10

BRAND% AG% GENERIC% ---- Inclusive of all drugs ---

Rep

ort

Per

cen

tage

Generic Entry

15

Lamotrigine – Labeled Events

Segmented Regression

Full Segmented

Regression Model*

(unadjusted for

prescription sales)

Total US Adverse Events US Serious Adverse Events

FDA

time

EVENT

time

MFR

time

FDA

time

EVENT

time

MFR

timeIntercept β0 147

(P= 0.71)

524

(P= 0.01)

525

(P=0.005)

41

(P=0.46)

88

(P=0.03)

90

(P=0.17)Trend before AG entered into market

(Period 1) β1

87

(P=0. 37)

22

(P= 0.64)

15

(P= 0.07)

17

(P=0.08)

5

(P=0.46)

7

(P=0.72)Level change after AG but before other

IG entered into market (Period 2) β2

-139

(P= 0.71)

-166

(P= 0.39)

-62

(P= 0.72)

-18

(P=0.62)

-18

(P=0.52)

19

(P=0.81)Trend change after AG but before other

IG entered into market (Period 2) β3

-80

(P= 0.51)

11

(P= 0.87)

-17

(P= 0.83)

-22

(P=0.06)

-5

(P=0.56)

-13

(P=0.63)Level change after other IG entered into

market (Period 3) β4

27

(P= 0.92)

-197

(P= 0.12)

-112

(P= 0.40)

9

(P=0.73)

-2

(P=0.90)

-1

(P=0.99)Trend change after other IG entered into

market (Period 3) β5

-12

(P= 0.89)

-38

(P= 0.42)

-1

(P= 0.99)

7

(P=0.42)

0.48

(P=0.94)

8

(P=0.68)

16Lamotrigine

Methodological Challenges

• Limited Authorized Generic Formulations

– Only specific formulations available as AG

• Lamotrigine chewable

• Extended release carbamazepine, metoprolol, venlafaxine

• Authorized Generic Availability

– Strength of connection between brand and AG

• Greenstone and Pfizer

– Dual marketing of AG and ANDA-approved generic

• Dr. Reddy’s Simvastatin

17

Methodological Approach

• Illustrated with alendronate

– Which report types?

– Which time?

– Ignore the AG?

• Illustrated with carbamazepine XR

– Multiple formulations with and without AG?

18

Sensitivity Analysis: Report Type?

19Alendronate – Gastritis

Sensitivity Analysis: Count All Time?

20Alendronate

Sensitivity Analysis: Ignore AG?

21Alendronate

Sensitivity Analysis: Modified Formulations?

22Carbamazepine

23

Brand

Authorized Generic

Generic

Sensitivity Analysis:

• FDA_date• Event_date• Report_date

Alendronate

0

500

1000

1500

2000

2500

3000

3500

4000

2 0 0 4 2 0 0 5 2 0 0 6 2 0 0 7 2 0 0 8 2 0 0 9 2 0 1 0 2 0 1 1 2 0 1 2 2 0 1 3 2 0 1 4 2 0 1 5

RE

PO

RT

S

YEAR

FDA_Time EVENT_Time MFR_Time

AG/IG Entry

0

50

100

150

200

2 0 0 4 2 0 0 5 2 0 0 6 2 0 0 7 2 0 0 8 2 0 0 9 2 0 1 0 2 0 1 1 2 0 1 2 2 0 1 3 2 0 1 4 2 0 1 5

RE

PO

RT

S

YEAR

FDA_Time EVENT_Time MFR_Time

AG/IG Entry

0

200

400

600

800

1000

1200

1400

1600

2 0 0 4 2 0 0 5 2 0 0 6 2 0 0 7 2 0 0 8 2 0 0 9 2 0 1 0 2 0 1 1 2 0 1 2 2 0 1 3 2 0 1 4 2 0 1 5

RE

PO

RT

S

YEAR

FDA_Time EVENT_Time MFR_Time

AG/IG Entry

Suicide / Suicidal Ideation in FAERS

24

Suicide / Suicidal Ideation in SHP

25

Sertraline Gabapentin Methylphenidate Zolpidem

BRAND GENERIC BRAND GENERIC BRAND GENERIC BRAND GENERIC

SAMPLE SIZE 7468 19015 3330 20597 6169 1630 7678 13215

Event count

(n, %)

35

(0.5%)

249

(1.3%)

14

(0.4%)

103

(0.5%)

25

(0.4%)

16

(1.0%)

39

(0.5%)

123

(0.9%)

Unadjusted p-

value<.0001 Ref 0.6873 Ref 0.0106 Ref 0.0007 Ref

Hazard Ratio

(95% CI)

0.53

(0.3- 0.8)Ref

1.07

(0.5- 2.4)Ref

0.37

(0.1- 1.0)Ref

0.85

(0.5- 1.4)Ref

Adjusted p-

value0.0031 Ref 0.8666 Ref 0.0506 Ref 0.5279 Ref

Specific Aim 3

• To develop a pilot surveillance system to compare patient experience for authorized generics and independent generics with brand name drugs for differential adverse event signal detection.

26

Marshfield’s Population

Security Health Plan MC Primary Service Area

Analysis Decision Points• Analysis cohort (e.g. first-time users of the drug vs. switch cohort)

• Maximum length of observation time (e.g. 365 days vs. 30 days)

• Inclusion / Exclusion of those with events prior to drug exposure

• Inclusion / Exclusion of those with events documented on a single date

• Statistical model (e.g. PH model of time to first event, negative binomial)

• Covariates

• Gender

• Age

• Charlson comorbidity score

• Diabetes indicator

• Smoking indicator

• MESA residency indicator

• Estimated SHP rate for the event of interest

• Review the raw data

• Initial model based analysis

• Are numbers small?

• Review unadjusted/adjusted significance

• Are hazard ratios reasonable?

Initial screening

• Evaluate results after varying decision points

• Evaluate subsetting cohort to limit date of first exposure

• Gather extensive data pre-exposure

• Evaluate associations with B/G and outcome

• Develop propensity score to improve B/G comparability

• Use another drug as secondary comparison group

Secondary analysis

• Trained research coordinator review

• Expert evaluationExpert Review

Surveillance System Development Process

30

Contact Information

Richard Hansen, RPh, PhDrah0019@auburn.edu

Auburn University

Harrison School of Pharmacy

Department of Health Outcomes Research & Policy

31