Table 1.
Descriptions of included studies and their reported outcomes
| Author (year), country | Study design | No of patients (mean age ± SD)
|
Patients’ diagnosis or disease condition | Method of payment | Method of medication adherence assessment | Primary outcomes (adherence rate) | Secondary outcomes | |
|---|---|---|---|---|---|---|---|---|
| Group 1 | Group 2 | |||||||
| Aarnio et al17 (2014), Finland | Retrospective cohort | 247,051 (61.0±8.17) | No controlled group | New stain user | Out-of-pockets | PDC | 54.6% | Higher out-of-pockets for statin associated with decrease medication in adherence (OR 0.80 per additional €0.10; 95% CI 0.80–0.80) The proportion of patients with good adherence reduced over time: 1-year (55%) vs 3-year follow-up (47%) |
| Castaldi et al16 (2010), USA | Longitudinal study | 13,891 (74) | No controlled group | Chronic pulmonary disease | Full Medicaid benefit Without Medicaid living in high-poverty neighborhoods Without Medicaid residing outside of high-poverty neighborhoods |
CRN | 69% | Patients who pay >USD20 per month out-of-pocket cost had higher CRN rate The odds of patients using inhalers would report CRN compared to patients who were not using inhalers (OR 1.43, 95% CI 1.212–1.69) |
| Sedjo et al21 (2011), USA | Retrospective cohort | 13,593 | No controlled group | Patients with breast cancer | Employer-based insurance scheme Noninstitutionalized insurance scheme Commercially insured insurance scheme |
MPR | 77% | Patients who pay out-of-pocket medication cost of $USD30 were more likely to be nonadherent compared to patients who pay <USD10 Younger age and heart condition comorbidities were associated with more nonadherent Within commercially insured population, the odds of nonadherence was increased among patients with higher out-of-pocket medication expenditure |
| William et al34 (2013), USA | Cross-sectional | 1,264 (74.6±5.6) | No controlled group | Patients with diabetes | Out-of-pockets money (exceed the limit of insurance coverage) | CRN | 84% | Patients who pay high out-of-pocket medication cost were more likely to report CRN Older age and annual income of <25 K were associated with high CRN |
| Zivin et al20 (2010), USA | Retrospective cohort | 3,071 | No controlled group | Chronic disease patients | Prescription drug coverage No drug coverage |
CRN | 80% | CRN was high among patients with no drug coverage and high out-of-pocket expenditure Lower household income, younger age, female, chronic condition, reduced self-rated health, depression, and reported adverse effect of medical treatment were more likely to report nonadherence |
| Zheng et al9 (2012), USA | Cross-sectional | 60 | No controlled group | Not mentioned | Private insurance Government subsidies No coverage |
CRN | No drug coverage: 28.6% Government subsidies: 88.9% Private insurance: 94.3% |
Patients who pay >USD100 out-of-pockets cost reported more CRN compared to patients who pay <USD20 |
| Dusetzina et al15 (2014), USA | Retrospective cohort | 1,541 (48.8±1) | No controlled group | Chronic myeloid leukemia | Coinsurance Copayment |
PDC | High copayment: 70% Low copayment: 79% |
|
| Ngo-Metzger et al19 (2011), USA | Cross-sectional | 1,135 | No controlled group | Patients with diabetes | Commercial insurance Medicare Medicaid No insurance |
CRN | Mexican American: 46.8% White: 72.8% Vietnamese: 72.4% |
Patients with high out-of-pocket medication cost and no drug coverage were more likely to report CRN Patients with low annual income, perceived financial barrier, and perceived financial burden were more likely to report CRN |
| Gibson et al24 (2010), USA | Cross-sectional | 96,734 (52.2±7.82) | No controlled group | Patients with diabetes | Comprehensive HMO PPO Point of service |
PDC | DM with OAD: 72.70% DM with OAD ± insulin: 74.5% |
Higher cost-sharing associated with reduced adherence Female, younger age, and less income more likely were associated with nonadherence |
| Law et al26 (2012), Canada | Cross-sectional | 5,732 | No controlled group | Chronic diseases | Insurance No insurance |
CRN | 90.4% | Patients with lack of insurance had more than a fourfold increased odds of CRN Patients with poor health, age <65 years, two or more chronic conditions, and lower income were more likely to report CRN |
| Kennedy et al28 (2011), USA | Cross-sectional | 8,935 | No controlled group | Not mentioned | Continuously insured Newly insured Continuously uninsured |
CRN | Reduction of 17.8% of reported CRN for newly insured patients who had got drug coverage through Part D | Younger age, multiple chronic diseases, depression, and poor health were likely to report CRN Newly insured patients were more likely to have resolved CRN (AOR 1.7, 95% CI 1.3–2.2) Reduction of 3.6% of reported CRN among continuously insured patients Reduction by 3.1% of reported CRN among continuously uninsured patients |
| Levine et al27 (2013), USA | Cross-sectional | 8,673 | No controlled group | Stroke | Medicare Private insurance Uninsured |
CRN | 88.6% | CRN increased among uninsured patients by 43.1% in 1999–2005 and 57.1% in 2006–2010 CRN increased among those with private insurance by 8.6% in 1999–2005 and 13.5% in 2006–2010 CRN remains stable for patients with Medicare insurance |
| Kim et al18 (2010), USA | Cross-sectional | 351 | No controlled group | Patients with Schizophrenia | VA insurance schemea Managed care (no VA, no Medicare, and no Medicaid)b Medicare/Medicaid (no managed care) Managed care and Medicare/Medicaid |
MMAS | Copayment burden: 28.5% Without copayment burden: 46.7% |
Patients with VA insurance scheme had negative association with complete adherence A negative relationship between copayment burden and complete rates of adherence (MMAS 4 items score of 0) |
| Maciejewski et al25 (2010), USA | Pre–post cohort | 60,017 | No controlled group | Diabetes, hypertension, and hyperlipidemia | Veterans Affairs exempt from medication copayment Veterans Affairs nonexempt from medication copayment |
PDC | Diabetes • Nonexempt 59.5% • Exempt 68.8% Hypertension • Nonexempt 56.0% • Exempt 56.7% Statin • Nonexempt 80.3% • Exempt 79.5% |
Adherence to antihypertensive medications was increased among exempted and nonexempted patients after the copayment increase (4.1% vs 5.9%) Adherence to statins was increased for exempted and nonexempted patients after the copayment increased (3.5% vs 6.6%), and adherence continued to increase (by 1.2%) for exempted veterans but decreased among nonexempted patients |
| Batavia et al30 (2010), India | Not mentioned | Tier 1: 156 (36) Tier 2: 141 (39) Tier 3: 242 (38) Tier 4: 96 (37) |
No controlled group | Patients with HIV | Tier 1: Receive first-line ART at no cost Tiers 2: Pay 50% Tier 3: Pay 75% Tier 4: Pay 100% of the respective cost |
Self-reported 3-day dose recall | Tier 1: 84.6% Tier 2: 71.6% Tier 3: 72.3% Tier 4: 79.2% |
No significant association was found between higher rate of optimal adherence and participant’s sex or marital status |
| Sears et al31 (2010), USA | Retrospective cohort | 7,858 | No controlled group | Overactive bladder | Free | MPR | 34% | 35.1% of patients who get medication at no charge did not refill their prescription Men were more adherent than women |
| Bhardwaja et al29 (2009), USA | Retrospective cohort | 269 | No controlled group | End-stage renal disease using Sevelamer |
Group 1: Medicare with annual cap Group 2: Medicare without annual cap on brand prescription drug spending |
PDC | Year 2003 • Group 1: 40% • Group 2: 66.3% Year 2004 • Group 1: 40.4% • Group 2: 59.2% |
Adherence rate reduced by 7.1% among group 2 and increased by 0.4% among group 1 |
| Donohue et al22 (2011), USA | Observational claim based (with controlled group) | No coverage: 1,877 Cap USD150: 1,069 Cap USD350: 8,395 |
Employer-sponsored: 3,739 | Patients with depression | No coverage Cap USD150 Cap USD350 Employer-sponsored |
MPR | Year 2004 • No drug coverage: 37% • USD150 cap: 37% • USD350 cap: 38% • Employer-sponsored: 51% Year 2005 • No drug coverage: 53% • USD150 cap: 51% • USD350 cap: 43% • Employer-sponsored: 48% |
Patients without limited coverage had higher rate of adherence compared to employer-sponsored group |
| Frankenfield et al32 (2011), USA | Cross-sectional | 216,127 | End-stage renal disease | Medicare beneficiaries | CRN | 77% | Smokers and chronic disease more likely to report CRN | |
| Harrold et al40 (2013), USA | Cross-sectional | Non-RA: 1,180 1–2 non-RA comorbid conditions: 6,275 ≥3 non-RA comorbid conditions: 6,824 |
RA: 219 | RA | Partial Full private Full public |
CRN | Patient with RA condition: 89.2% Patient with non-RA condition: 46% Patient with 1–2 non-RA conditions: 59.5% Patient with ≥3 non-RA conditions: 64.5% |
Partial coverage CRN (OR 0.91, 95% CI 0.81–1.03), full private coverage CRN (OR 0.52, 95% CI 0.47–0.57), and full public coverage CRN (OR 0.82, 95% CI 0.75–0.90) |
| Zivin et al23 (2009), USA | Cross-sectional | 24,234 | No controlled group | Patients with depression | No coverage Partial coverage Employer coverage Medicaid Part D |
CRN | Depression • 2004: 73% • 2005: 73% • 2006: 76% Without depression • 2004: 87% • 2005: 88% • 2006: 91% |
Patients without depression had better adherence rate compared to patient with depression |
Notes:
a
VA is a US cabinet department that provides patient care, veterans’ benefits, and other services to the US armed forces and their family.
b
Managed care is a type of health insurance that have contracts with health care providers and medical facilities to provide care for members at reduced cost.
Abbreviations: SD, standard deviation; PDC, proportion of days covered; OR, odds ratio; CI, confidence interval; CRN, cost-related nonadherence; MPR, medication possession ratio; DM, diabetes mellitus; OAD, oral antidiabetic drug; HMO, Health Maintenance Organization; AOR, adjusted odds ratio; VA, Veterans Administration; MMAS, Morisky Medication Adherence Scale; ART, antiretroviral therapy; RA, rheumatoid arthritis; PPO, Preferred Provider Organization.