INDIVIDUAL AND SOCIETAL BURDEN OF NON-ADHERENCETO ANTIDEPRESSANTS IN BRAZIL

Riccardo Pedersini1, Natalia Flores2, Alvaro Mitsunori Nishikawa3

PMH9

Poster Presented at the ISPOR 5th Latin America Conference 6 -8 September 2015 Santiago, Chile ©Copyright 2015 Evidências Credibilidade Científica, A Kantar Health Company Alameda Lorena, 131, Conj. 115 e 117, Jardim Paulista São Paulo, SP - 01424-001 Brazil + 55-11-3884-0942

www.evidencias.com.br www.kantarhealth.com

1Kantar Health, Epsom UK; 2Kantar Health, Foster City, CA USA; 3Evidências _ Kantar Health, São Paulo, Brazil

Depression is a serious public health problem related to high social costs and risk of suicide, also the leading cause of disability affecting 350 million people worldwide according to the World Health Organization (WHO).1,2

Although efficacious and cost-effective treatments are available, non-adherence is very common and is one of the most significant issues for successful treatment.

WHO recognizes depression as one of 9 chronic conditions to focus on for improvingmedication adherence.3 About 1/3 of patients discontinue treatment within the 1st month, and about 45% do not exceed the 3rd month of treatment.4,5

Odds of a good outcome in medication-adherent patients are almost three times higher than in low adherence ones.6 Studies show that poor adherence leads to quality of life decrements, mental health deterioration and increased expenditure, mainly due to indirect costs.6-8

There is a lack of data on this issue in Brazil, so this study investigates the associationbetween adherence and the burden of depression in the Brazilian population.

INTRODUCTION

•

•

•

•

•

Non-adherence with antidepressant medication is a frequent problem for managing depression.

While depression itself can be associated with negative health outcomes,1,2 this studyprovides support that the impact can be further exacerbated by non-adherence to medication.

In this real-world study of Brazilian patients with depression, non-adherence demonstrated a negative impact on patients’ outcomes, specifically to medication satisfaction, quality of life, and work productivity, indicating that medication adherence is an important factor in treatment success.

Further research needs to be conducted to help determine factors leading to non-adherence and identify interventions so adherence can be improved and negative impact lessened.

•

•

•

•

CONCLUSION

Non-adherence was significantly associated with lower educational level, working full time,lower income, and public insurance coverage.

Additionally, non-adherence was associated with lower mental health-related quality of life andlower health utilities.

Lastly, non-adherence was associated with greater presenteeism, which can indicate higher indirect costs.

•

•

•

DISCUSSION

Presenteeism measured by the WPAI was higher for the non-adherent group (42 vs. 37, see Figure 3). No other significant differences for WPAI items were found.

•

All results held significant after controlling for covariates (socio-demographic and health characteristics) using GLM models, apart from differences in hospitalization.

•

Table 1. Socio-demographic and Health Characteristics

Chi-2 / t-test

N Percent N Percent N Percent P*PHQ-9 score 0.007*Mean, (SD) 7.40 (6.98) 7.15 (6.92) 8.33 (7.16)

PHQ-9 severity 0.031*Minimal 716 48.2% 587 50.2% 129 40.7%

Mild 378 25.4% 293 25.0% 85 26.8%Moderate 131 8.8% 96 8.2% 35 11.0%

Moderately severe 125 8.4% 93 7.9% 32 10.1%Severe 137 9.2% 101 8.6% 36 11.4%

Max Rx satisfaction (1-7) <0.001*Mean, (SD) 5.17 (1.52) 5.25 (1.52) 4.86 (1.47)

Age <0.001*Mean, (SD) 40.32 (12.42) 40.98 (12.55) 37.86 (11.61)

Gender 0.199Female 1040 69.9% 809 69.1% 231 72.9%Male 361 30.9% 86 27.1% 447 30.1%

Education level 0.015*Less than high school 138 9.3% 99 8.5% 39 12.3%

High school 698 46.9% 539 46.1% 159 50.2%University degree 651 43.8% 532 45.5% 119 37.5%

Working full/part time 0.049*No 549 36.9% 447 38.2% 102 32.2%Yes 938 63.1% 723 61.8% 215 67.8%

Annual household income 0.041*Up to R$ 1,000 145 9.8% 111 9.5% 34 10.7%

R$ 1,001 to 3,000 383 25.8% 283 24.2% 100 31.5%R$ 3,001 to 5,000 214 14.4% 176 15.0% 38 12.0%

R$ 5,001 to 10,000 238 16.0% 183 15.6% 55 17.4%More than R$ 10,000 342 23.0% 282 24.1% 60 18.9%

Decline to answer 165 11.1% 135 11.5% 30 9.5%Married/living with partner 0.303

No 577 49.3% 146 46.1% 723 48.6%Yes 764 51.4% 593 50.7% 171 53.9%BMI 0.197

Mean, (SD) 27.16 (5.95) 27.27 (6.01) 26.78 (5.74)CCI 0.374

Mean, (SD) 0.70 (1.53) 0.72 (1.60) 0.63 (1.23)Private insurance 0.025*

No 490 41.9% 155 48.9% 645 43.4%Yes 842 56.6% 680 58.1% 162 51.1%

Visited PCP 0.028*No 440 37.6% 98 30.9% 538 36.2%Yes 949 63.8% 730 62.4% 219 69.1%

Visited specialist 0.170No 856 73.2% 244 77.0% 1100 74.0%Yes 387 26.0% 314 26.8% 73 23.0%

Hospitalized 0.001*No 952 81.4% 231 72.9% 1183 79.6%Yes 304 20.4% 218 18.6% 86 27.1%

Visited ER 0.066No 651 55.6% 158 49.8% 809 54.4%Yes 678 45.6% 519 44.4% 159 50.2%

0.351No 772 66.0% 218 68.8% 990 66.6%Yes 497 33.4% 398 34.0% 99 31.2%

Figure 2. Quality of life (SF-36v2)

RESULTS

The projection of 1,457 respondents to the 2012 Brazilian population resulted in 14.82 M people taking prescription anti-depressants (7.5% of the total population), of which 75% were classified as adherent and 25% as non-adherent. These proportions are wellrepresented in the sample (79% vs. 21%).

Most respondents were female (70%), and the mean age was 41 years for the adherent group and 38 years for non-adherent ones. Adherents were more educated, had higher household income, insurance and less of them were employed (see Table 1).

Non-adherent respondents reported significantly more severe depression (22% vs. 17% with PHQ-9 score ≥ 15) compared to adherent ones, and PHQ-9 scores grew consistently with MMAS-4 scores (see Table 1 and Figure 1).

Non-adherent respondents were significantly less satisfied with their antidepressant medication (4.9 vs. 5.3, see Table 1).

•

•

•

•

ReferencesMarcus M, Yasamy MT, van Ommeren M, Chisholm D, Saxena S. Depression. A Global Public Health Concern. 2012; http://www.who.int/mental_health/management/depression/who_paper_depression_wfmh_2012.pdf. Accessed July 20th, 2015.Investing in Mental Health. 2003; http://www.who.int/mental_health/media/investing_mnh.pdf. Accessed July 10th, 2015.Adherence to LongTerm Therapies: Evidence for Action. 2003.Bultman DC, Svarstad BL. Effects of phisician communication style on client medication beliefs and adherence with antidepressant treatment. Patient Education and Counseling. 2000;40:173-185.Nemeroff CB. Improving antidepressant adherence. J Clin Psychiatry. 2003;64 Suppl 18:25-30.DiMatteo MR, Giordani PJ, Lepper HS, Croghan TW. Patient adherence and medical treatment outcomes: a meta-analysis. Med Care. 2002;40(9):794-811.Stein-Shvachman I, Karpas DS, Werner P. Depression Treatment Non-adherence and its Psychosocial Predictors: Differences between Young and Older Adults? Aging Dis. 2013;4(6):329-336.Cantrell CR, Eaddy MT, Shah MB, Regan TS, Sokol MC. Methods for evaluating patient adherence to antidepressant therapy: a real-world comparison of adherence and economic outcomes. Med Care. 2006;44(4):300-303.Kroenke K, Spitzer RL, Williams JB. The PHQ-9. Journal of general internal medicine. 2001;16(9):606-613.Morisky DE, Green LW, Levine DM. Concurrent and predictive validity of a self-reported measure of medication adherence. Medical care. 1986;24(1):67-74.Ware J, Kosinski M, Turner-Bowker D, Gandek B. User’s manual for the SF-12v2. Health Survey with a supplement documenting SF-12® Health SurveyQualityMetric Incorporated, Lincoln, RI. 2002.Brazier J, Roberts J, Deverill M. The estimation of a preference-based measure of health from the SF-36. Journal of health economics. 2002;21(2):271-292.Reilly MC, Zbrozek AS, Dukes EM. The validity and reproducibility of a work productivity and activity impairment instrument. Pharmacoeconomics. 1993;4(5):353-365.

METHODOLOGY

•

Data from the 2011 & 2012 Brazil National Health and Wellness Survey (NHWS; N=24,000), a self-administered, Internet-based general health questionnaire from a sample of adults (18+), supplemented by CATI (computer assisted telephone interviewing) to reach those without internet access.

Respondents recruited from an internet panel using a random stratified samplingframework to ensure the demographic composition (with respect to age and gender) isrepresentative of the adult Brazilian population based on the data from International Data Base of the U.S. Census Bureau and Organization for Economic Cooperation and Development.

Of the 24,000 total NHWS respondents, 22,744 had a complete record of responses (94%).

Of the latter, 2,721 (12%) reported a diagnosis of depression and 1,457 (6%) had aprescription medication for depression (Rx). Those with an Rx constituted our sample.

Data source

Sample

Measures

Demographics and health characteristics

Depression severity

Medication Adherence

Outcome measures

•

•

•

•

•

•

•

•

•

•

•

Age, gender, marital status, education, household income, insurance type.

Body mass index (BMI), smoking status, alcohol use, exercise behavior. The Charlsoncomorbidity index (CCI) was also examined for general comorbidity burden.

Patient Health Questionnaire-9 (PHQ-9) to assess severity of depression (Minimal, Mild, Moderate, Moderately Severe, Severe) in respondents with a physiciandiagnosis of depression.9

Medication Status: respondents were asked if they were currently taking a prescription medication for depression.

Morisky Medication Adherence Scale-4 (MMAS-4)10 was used to categorize those with low adherence (score ≥ 3) as “non-adherent” and the others (medium/highadherence, score ≤ 2) as “adherent.”

Health-related quality of life: HRQoL measured via Medical Outcomes Study Short Form (SF-12v2 for 2011 NHWS; SF36v2 for 2012 NHWS)11,12 including: Mental Component Summary (MCS), Physical Component Summary (PCS) and Health Utilities (SF-6D)

Work productivity loss: measured via the Work Productivity and Activity Impairment-General Health scale (WPAI-GH)13 with Absenteeism: (% time missed from work due to health problems), Presenteeism (% time impaired while working), Overall Work Impairment (% time overall work productivity impaired as a result of absenteeism and presenteeism) and Activity Impairment (% time health problems affected regular activities)

Healthcare resource use: frequency of various forms of healthcare resource use for the past 6 months, visits to Primary care physician (PCP), Specialist (psychiatrist or psychologist) and Emergency room (ER), and Hospitalization

Adherent respondents (according to MMAS-4) were compared to the non-adherent onseverity (PHQ-9), socio-demographics, health characteristics, health-related quality of life (SF-12 or SF-36), work productivity and activity impairment (WPAI) and healthcare resource use (physician, hospital and emergency visits) using chi-squared tests for categorical variables and t-tests for cardinal variables.

A general linear modeling (GLM) approach was used to test the association of adherence with depression and other outcomes (HRQoL, WPAI, resource use) while controlling forcovariates (socio-demographics and health characteristics).

Statistical analyses•

•

Note: *Statistically significant (p < 0.05) differences among adherent and non-adherent respondents are underlined

Total(N = 1,487)

Adherent(N = 1,170)

Non-adherent(N = 317)

50

70

60

40

30

20

10

0MCS PCS

*

*

Health u�lity

MCS

/ P

CS s

core

Adherent Non-adherent

Health-related quality of life(SF-36v2): Mental componentsummary (MCS), Physical component summary (PCS) and Health utility score for adherent (yellow) andnon-adherent (red) respondents.Error bars: standard error of the mean. Starred differences: p < 0.05.

Figure 3. Work productivity loss (WPAI)

Figure 4. Healthcare resource use

50

60

40

30

20

10

0Absenteeism Presenteeism Overall work

impairmentAc�vity

impairment

*

Perc

ent

Adherent Non-adherent

2,0

2,5

1,5

1,0

0,5

0PCP Specialist Hospital Emergency

N.v

isits

(pas

t 6 m

onth

s)

Adherent Non-adherent

Work productivity and activityimpairment (WPAI) foradherent (yellow) andnon-adherent (red) respondents. Error bars: standard error of the mean. Starred differences: p < 0.05.

Healthcare resource utilization in past 6 months: number of visits to a primary care physician (PCP), Specialist (psychiatrist or psychologist),Hospital and Emergency for adherent (yellow) and non-adherent (red)respondents. Error bars: standarderror of the mean. No significantdifference was found.

Non-adherence showed effects in terms of burden of disease. The Mental Component Summary of the SF-12 or SF-36v2 measuring quality of life was lower for the non-adherent group (MCS: 33 vs. 36) and also had lower Health Utility score (.59 vs. .60, see Figure 2)

•

No significant difference was detected in healthcare resource utilization after controlling for covariates(see Figure 4).

•

Figure 1. Depression severity (PHQ-9)

Severity of depression measured by the PHQ-9 (score: 0-27) against adherence to antidepressantsmeasured by MMAS-4 (score: 0-4). Non-adherent in yellow; adher-ent in red; 3rd-degree polynomial regression line in grey; error bars: standard error of the mean; starred difference: p < 0.01.

9.5

9.0

8.5

8.0

7.5

7.0

6.5

6.0

MMAS-4 score Adherent Non-adherent

PHQ

-9 s

core

0 1 2 3 4

*

Currently participate in talk therapy

1.

2.3.4.

5.6.7.

8.

9.10.11.

12.13.