F2: HOW DOES MEASUREMENT OF MULTIPLE MEDICATION ADHERENCE DIFFER - - PDF document

1

F2: HOW DOES MEASUREMENT OF MULTIPLE MEDICATION ADHERENCE DIFFER BETWEEN CHRONIC DISEASES?

An ISPOR Forum by the Multiple Medication Adherence Measurement Working Group of the ISPOR Medication Adherence and Persistence Special Interest Group Monday, May 22, 2017

HOW DOES MEASUREMENT OF MULTIPLE MEDICATION ADHERENCE DIFFER BETWEEN CHRONIC DISEASES?

Mia Malmenäs, MSc, Director, Real World Strategy & Analytics, Mapi, Stockholm, Sweden Priti Pednekar, MPharm, Graduate Student (PhD), Mayes College

• f Healthcare Business and Policy, University of the Sciences,

Philadelphia, PA, USA Bijan J. Borah, PhD, Associate Professor of Health Services Research, Director/Lead Consultant, Economic Evaluation Service, Kern Center for the Science of Healthcare Delivery, Mayo Clinic College of Medicine, Rochester, MN, USA Bryan Bennett, Director, Patient Centered Outcomes, Adelphi Values, Adelphi Mill, UK

2

2

MULTIPLE MEDICATION ADHERENCE MEASUREMENT WORKING GROUP

Chairs: Mia Malmenäs, MSc, Director, Real World Strategy & Analytics, Mapi, Stockholm, Sweden Andrew M. Peterson, PharmD, PhD, Dean, Mayes College of Healthcare Business and Policy, University of the Sciences, Philadelphia, USA Leadership: Lusine Abrahamyan, MD, MPH, PhD, THETA, Toronto, Canada Tamás Ágh, MD, MSc, PhD, Principal Researcher, Syreon Research Institute, Budapest, Hungary Bryan Bennett, Director, Patient Centered Outcomes, Adelphi Values, Adelphi Mill, UK Bijan J. Borah, PhD, Associate Professor, Mayo Clinic College of Medicine, Division

• f Health Care Policy and Research, Rochester, MN, USA

Thomas J. Bunz, Pharm.D., Ph.D, Director of Analytic Strategy, Aetna Consumer Analytics, Hartford, Conn., USA

3

MULTIPLE MEDICATION ADHERENCE MEASUREMENT WORKING GROUP

Leadership: Mickaël Hiligsmann, PhD, Assistant Professor, Department of Health Services Research of Maastricht University, the Netherlands David Hutchins, MHSA, MBA, Senior Advisor, Strategic Research, CVS Caremark, USA Elizabeth Manias, PhD, University of Melbourne, Australia Priti Pednekar, M.Pharm, B. Pharm, Graduate Student, Mayes College of Healthcare Business and Policy, University of the Sciences, PA, USA Amit Raval, M.Pharm, PhD, Senior Researcher, Healthcore, Inc., Wilmington, DE, USA Adina Turcu-Stiolica, PharmD, PhD, Chair of Pharmaceutical Marketing and Management, University of Medicine and Pharmacy Craiova, Romania Allison F. Williams, RN, PhD, School of Nursing and Midwifery, Monash University, Victoria, Australia John E. Zeber, PhD, Central Texas Veterans Health Care System, Temple, TX ; Scott & White Healthcare, Center for Applied Health Research, Temple, TX, USA

4

3

SPECIAL INTEREST GROUP OBJECTIVE

5

Conduct a systematic literature search of methods used calculating medication adherence for multiple drugs Summarize the results overall as well as by disease area Investing the accuracy of the measurements by conducting simulation Evaluate the measurements using real data Provide recommendations on measurements for evaluation of medication adherence in patients using multiple drugs

SPECIAL INTEREST GROUP OUTCOMES Two workshops One podium presentation Two publications, one submitted and one in process Today’s Forum

6

4

Is it necessary to have unique multiple adherence measurements for different diseases?

7

Is there an ideal multiple adherence measurement available today?

8

5

Bijan J. Borah, PhD, Associate Professor of Health Services Research, Director/Lead Consultant, Economic Evaluation Service, Kern Center for the Science of Healthcare Delivery, Mayo Clinic College of Medicine, Rochester, MN, USA

9

FLOW DIAGRAM OF THE SYSTEMATIC LITERATURE REVIEW PROCESS

Included Eligibility Screening Identification

Records identified through database searching (n = 1,706) Records after duplicates removed (n = 1,382) Records screened (n = 1,382) Full-text articles assessed for eligibility (n = 301) Studies included in qualitative synthesis (n = 151) Records excluded (n = 1,081) Additional records identified through

• ther sources

(n = 8) Full-text articles excluded, with reasons (n = 150)

• Not original research (n = 9)
• No evaluation method for medication

adherence (n= 52)

• Studies not assessing multiple medication

adherence (n= 83)

• Studies assessing adherence to guidelines

but not medications (n= 1)

• Studies assessing adherence to diet (n = 2)
• Other (n= 22)

Based on the PRISMA template: Panic, N., et al., PLoS One, 2013. 8(12): e83138. 18 10

6

DISTRIBUTION OF STUDIES BY STUDY DESIGNS

11

130 20 1 20 40 60 80 100 120 140 Observational Studies Randomized Controlled Trial Validation study

Cross- sectional; 25; 19% Prospectiv e Cohort; 47; 36% Retrospecti ve Cohort; 58; 45%

NUMBER OF MMA METHODS USED BY STUDY DESIGN

12

Number of MMA Methods Used Research Design 1 2 3 4 Total Observational Cross- sectional 20 1 4 25 Observational Prospective 33 10 3 1 47 Observational Retrospective 43 8 6 1 58 RCT 14 3 2 1 20 Validation Study 1 1 Total 110 23 15 3 151

7

TYPES OF DATA COLLECTED BY STUDY DESIGN

13 10 20 30 40 50 60 70 80 90 24 1 1 1 2 1 39 3 6 5 3 2 4 2 2 9 17 8 26 16 5 1 12 2 3 3 1 3 6 3 2 1 Validation Study RCT

• Obs. Retrospective
• Obs. Prospective
• Obs. Cross-sectional

SPECIFIC MMA METHODS BY STUDY DESIGN

14

10 20 30 40 50 60 70 80 90 2 1 1 1 24 1 1 2 2 4 4 1 39 2 25 16 13 9 3 9 1 1 2 4 1 12 1 3 1 Validation Study RCT

• Obs. Retrospective
• Obs. Prospective
• Obs. Cross-sectional

8

Priti Pednekar, MPharm, Graduate Student (PhD), Mayes College of Healthcare Business and Policy, University of the Sciences, Philadelphia, PA, USA

15

ADHERENCE MEASUREMENT METHODS BY DISEASE

16

0.0% 10.0% 20.0% 30.0% 40.0% 50.0% 60.0% 70.0% 80.0% 90.0% Cardivascular disorders Sexually transmitted disorders (HIV/AIDS) Metabolic disorders Mental disorders

Quasi-Objective Self-report method Therapeutic outcome monitoring Drug level monitoring

9

QUASI-OBJECTIVE METHODS (1)

17

0% 5% 10% 15% 20% 25% 30% 35% 40% 45% PDC MPR Medication gaps Other method Time to discontinuation Persistence rate Cardivascular disorders Sexually transmitted disorders (HIV/AIDS) Metabolic disorders Mental disorders

46% of studies used quasi-objective methods to measure MMA

QUASI-OBJECTIVE METHODS (2)

Total Number of Studies 46 PDC 19 41.3% MPR 10 21.7% Medication gaps 1 2.2% Other method for calculating adherence rate 2 4.3% Time to discontinuation 9 19.6% Persistence rate 6 13.0%

18

Cardiovascular Disorders Sexually Transmitted Disease (HIV/AIDS)

Total Number of Studies 45 PDC 3 6.7% MPR 1 2.2% Medication gaps 3 6.7% Other method for calculating adherence rate 8 17.8% Time to discontinuation 1 2.2% Persistence rate 1 2.2%

10

QUASI-OBJECTIVE METHODS (3)

Total Number of Studies 18 PDC 3 16.7% MPR 7 38.9% Medication gaps 1 5.6% Other method for calculating adherence rate 2 11.1% Time to discontinuation 3 16.7% Persistence rate 3 16.7%

19

Metabolic Disorders Mental Disorders

Total Number of Studies 15 PDC 3 20.0% MPR 3 20.0% Medication gaps 1 6.7% Other method for calculating adherence rate 2 13.3% Time to discontinuation 1 6.7% Persistence rate 1 6.7%

MPR FOR MMA (N=25)

Calculation method n % Diseases investigated Average of (∑days of supply for each medication/study period) 4 2.6% Cardiovascular (2), Metabolic (2) ∑days of supply for all medications/study period 3 2.0% Cardiovascular (3) ∑days with supply for any medication/study period 2 1.3% Mental (1), Urological (1) Average of (∑days of supply/(days between last prescription- first prescription) per medications); supply obtained in the last fill was excluded 2 1.3% Cardiovascular (1), Metabolic (1) (∑days of supply/(days between last prescription- first prescription) per medications) and if all were >80% then adherent; supply

• btained in the last fill was excluded

1 0.7% Cardiovascular (1), Mental (1), Metabolic (1), Musculoskeletal (1), Respiratory (1) (∑days of supply/days of eligibility for that medication) per medications and if all were ≥80% then adherent 1 0.7% Cardiovascular (1) ∑days of supply for all medications/(days between last prescription- first prescription + days of supply for last fill - number of days in hospital) 1 0.7% Neurological (1) ∑days of supply for mutliple medications/(days between last prescription- first prescription + days of supply for last fill) 1 0.7% STD (1) ∑tablets dispensed/∑tablets recommended or prescribed 1 0.7% Metabolic (1) Weighted average of (∑days for supply)/days for which medication was needed) per medications) 1 0.7% Mental (1)

20

11

Different Formulas of MPR for MMA

1 2 3 4 5 6 Cardiovascular Disorders Sexually transmitted disorders (HIV/AIDS) Metabolic Disorders Mental Disorders

21

10 Different formulas

• f MPR were used to

measure MMA n = 10 n = 1 n = 7 n = 3

n= Number of studies which used MPR to calculate MMA

PDC FOR MMA (N=29)

Calculation method n % Diseases investigated ∑days supplied for all medications/study period 13 8.6% Cardiovascular (9), Mental (2), STD (2), Metabolic (1) Average of ∑days supplied per medications/study period 4 2.6% Cardiovascular (3), Cancer (1), Musculoskeletal (1), Neurological (1) ∑days supplied for any medication/study period 3 2.0% Cardiovascular (2), Metabolic (1), Respiratory (1) Average of (∑days supplied/(days between last prescription- first prescription) + days supplied for the last fill)) per medications) 2 1.3% Cardiovascular (1), Metabolic (1) ∑days supplied for all medications/(study period - days of hospitalization) 1 0.7% STD (1) ∑days supplied/study period for each medication and if all were ≥80% then adherent 1 0.7% Cardiovascular (1) ∑days supplied/study period for each medication, adherent if at least 3 out 4 drugs were taken 50% during the study period 1 0.7% Cardiovascular (1) ∑days supplied/study period for each medication, adherent if drugs were taken 50% during the study period 1 0.7% Cardiovascular (1) ∑days for supplied per medications/(study period x number of medications) 1 0.7% Cardiovascular (1)

22

12

Different Formulas of PDC for MMA

1 2 3 4 5 6 7 8 9 10 Cardiovascular Disorders Sexually transmitted disorders (HIV/AIDS) Metabolic Disorders Mental Disorders

23

9 Different formulas of PDC were used to measure MMA n = 19 n = 3 n = 3 n = 3

n= Number of studies which used PDC to calculate MMA

MISSING DOSES/DAYS FOR MMA (N=6)

Calculation method n % Diseases investigated ∑days without medications/study period 3 2.0% Metabolic (1), STD (2), ∑days without medications/(days between last prescription- first prescription) 1 0.7% Cardiovascular (1) (1-∑doses of medications/∑expected doses of medications) x 100 1 0.7% STD (1) ∑days without medications 1 0.7% Mental (1)

24

13

OTHER METHODS FOR MMA (N=20)

Calculation method n % Diseases investigated Composite Adherence Score (CAS) = a hierarchical algorithm, which combines adherence data from MEMS, pill count, and self-report 3 2.0% STD (3) Continuous Multiple-interval Measures of Medication Availability (CMA) = the sum of the all of the days’ supply of medication / the number of days between the first fill and the last refill 1 0.7% Cardiovascular (1) Covered minutes per day = ((1440 min - uncovered min)/1440 min) x 100 1 0.7% STD (1) Daily Patient Possession Ratio (DPPR) = look at each day in the

• bservation period separately, and determine how many

medications are available, set a score between 0 (no medication available) and 1 (all medications available) weighted by the number

• f medications to be taken each day, resulting in daily scores

indicating the proportion of medications available for each day; sum the scores and divide by the number of days in the observation period to obtain the proportion of all medications available for daily use 1 0.7% NR (1) Medication total (MED TOT) = ∑supply of pills dispensed/number of days elapsed 1 0.7% Metabolic (1) Overall pill count adherence score represented the mean pill count adherence across all prescribed medications 1 0.7% Mental (1) Proportion of medications taken during the past week 1 0.7% STD (1)

25

Bryan Bennett, Director, Patient Centered Outcomes, Adelphi Values, Adelphi Mill, UK

26

14

SELF-REPORT METHODS (1)

Total Number of Studies 46 PRO questionnaire 12 26.1% MARS 2 4.3% MMAS-4 4 8.7% MMAS-8 6 13.0% MOS-SAS 1 2.2% Other self-report method (unnamed questionnaire, interview, undefined) 3 6.5% Informant rating (doctor/nurse report) 1 2.2%

27

Cardiovascular Disorders Sexually Transmitted Disease (HIV/AIDS)

Total Number of Studies 45 PRO questionnaire 19 42.2% ACTG [also includes modified version] 6 13.3% PMAQ [also includes revised version] 3 6.7% MATI 1 2.2% CPCRA 4 8.9% MAF 1 2.2% MARS 1 2.2% MMAS-4 2 4.4% RAMS 1 2.2% PAM 1 2.2% SHCS-AQ 2 4.4% Other self-report method (unnamed questionnaire, interview, undefined) 20 44.4% Informant rating (doctor/nurse report) 1 2.2%

SELF-REPORT METHODS (2)

28

Metabolic Disorders Mental Disorders

Total Number of Studies 15 PRO questionnaire 3 20.0% MARS 1 6.7% TRQ 1 6.7% Adherence Questionnaire 1 6.7% Other self-report method (unnamed questionnaire, interview, undefined) 2 13.3% Informant rating (doctor/nurse report) 2 13.3% Total Number of Studies 18 PRO questionnaire 7 38.8% MMAS-4 6 33.3% MMAS-8 2 11.1% SDSCA 1 5.6% Other self-report method 3 16.7% Informant rating 0.0%

15

SELF-REPORT METHODS (3)

29

0.0% 5.0% 10.0% 15.0% 20.0% 25.0% 30.0% 35.0% 40.0% 45.0% 50.0% PRO Other self-report Informant rating Cardiovascular disorder Sexually transmitted disease Metabolic disorder Mental disorder

59% of studies used self-report methods to measure MMA

Mia Malmenäs, MSc, Director, Real World Strategy & Analytics, Mapi Sweden AB, Stockholm, Sweden

30

16

WHY ARE DIFFERENT MEASURES USED?

• Purpose of the study (why adherence is being measured?)
• Type of study design
• Availability and accessibility of data
• Goal of the treatment/adherence program
• Treatment pattern changes
• Reliability of measures
• Practicality of measures
• Cost-effectiveness of measures
• Degree of complexity of calculation, analysis and interpretation
• Guidelines (eg; PQA)
• Existing knowledge

31

WHAT KIND OF MMAS HAS WORKED BEST FOR YOU?

32

17

33

DO YOU PREFER USING DIFFERENT MMAS FOR DIFFERENT DISEASES?

SUMMARY

• Methodological approaches for measuring MMA differ by study

design

• There is no single quasi-objective measure which is preferred

to calculate MMA for a particular disease. However, PDC and MPR were most commonly used quasi-objective measures

• Self-report measures are most commonly used in the literature
• Self-report measures need further research to examine their

validity in reliability in different disease settings

34

18

• Sign up as Review

Group Member

• Join ISPOR

Special Interest Groups

• Need ISPOR

membership number

• Business card to

Theresa or email

ttesoro@ispor.org

Sign up as Review Group Member

Don’t forget to submit your evaluation using the ISPOR APP