Too Good to Be True? Social Desirability Bias in Age-Related Patterns of Self-Reported Medication Adherence Among Patients With Atrial Fibrillation
Author(s)
Chen-Chia Pan, MA, Sabrina Müller, MSc.
GIPAM GmbH, Wismar, Germany.
GIPAM GmbH, Wismar, Germany.
OBJECTIVES: To investigate whether social desirability bias contributes to the paradox of older adults reporting high medication adherence despite a greater risk of non-adherence due to polypharmacy and comorbidities.
METHODS: We analyzed cross-sectional data from 308 anonymized adults with atrial fibrillation (AF), surveyed via online panels in the UK and the US, and during outpatient visits at a specialized stroke center in a German hospital. Patients suffering from AF and undertaking a long-term treatment were included independently of the type, treatment, or duration of the disease. Participants completed the Adherence Assessment Questionnaire (AAQ), a validated self-report instrument that includes a two-item social desirability scale. Binomial logistic regression was used to estimate associations between age, social desirability, and their interaction with self-reported non-adherence, adjusting for gender, country, and number of medications. Sensitivity analyses using ordinal and linear regression models, as well as moderation estimates, were conducted to assess robustness.
RESULTS: The sample had a mean age of 67.0 years (SD: 13.2), and 33.8% were female. Increasing age and higher social desirability scores were associated with lower odds of reporting non-adherence (OR = 0.97, 95% CI: 0.94-0.99, p = .009 and OR = 0.80, 95% CI: 0.64-1.00, p = .051, respectively). The age-by-social desirability interaction was marginal in the binomial model (OR = 0.98, 95% CI: 0.96-1.00, p = .105), but significant in ordinal (p < .001) and linear (p = .002) regressions. Older adults with higher social desirability scores were especially likely to report better adherence. The moderation analysis (p = .016) confirmed that social desirability bias strengthened the association between age and adherence. No country-level differences were observed.
CONCLUSIONS: The apparent increase in self-reported adherence with age may reflect response bias rather than true behavioral differences. The tendency of socially desirable response behavior should be considered when interpreting self-reported adherence data.
METHODS: We analyzed cross-sectional data from 308 anonymized adults with atrial fibrillation (AF), surveyed via online panels in the UK and the US, and during outpatient visits at a specialized stroke center in a German hospital. Patients suffering from AF and undertaking a long-term treatment were included independently of the type, treatment, or duration of the disease. Participants completed the Adherence Assessment Questionnaire (AAQ), a validated self-report instrument that includes a two-item social desirability scale. Binomial logistic regression was used to estimate associations between age, social desirability, and their interaction with self-reported non-adherence, adjusting for gender, country, and number of medications. Sensitivity analyses using ordinal and linear regression models, as well as moderation estimates, were conducted to assess robustness.
RESULTS: The sample had a mean age of 67.0 years (SD: 13.2), and 33.8% were female. Increasing age and higher social desirability scores were associated with lower odds of reporting non-adherence (OR = 0.97, 95% CI: 0.94-0.99, p = .009 and OR = 0.80, 95% CI: 0.64-1.00, p = .051, respectively). The age-by-social desirability interaction was marginal in the binomial model (OR = 0.98, 95% CI: 0.96-1.00, p = .105), but significant in ordinal (p < .001) and linear (p = .002) regressions. Older adults with higher social desirability scores were especially likely to report better adherence. The moderation analysis (p = .016) confirmed that social desirability bias strengthened the association between age and adherence. No country-level differences were observed.
CONCLUSIONS: The apparent increase in self-reported adherence with age may reflect response bias rather than true behavioral differences. The tendency of socially desirable response behavior should be considered when interpreting self-reported adherence data.
Conference/Value in Health Info
2025-11, ISPOR Europe 2025, Glasgow, Scotland
Value in Health, Volume 28, Issue S2
Code
PCR1
Topic
Methodological & Statistical Research, Patient-Centered Research, Study Approaches
Topic Subcategory
Adherence, Persistence, & Compliance, Patient Behavior and Incentives, Patient-reported Outcomes & Quality of Life Outcomes
Disease
Cardiovascular Disorders (including MI, Stroke, Circulatory)