Beliefs about Medicine Questionnaire (BMQ); BMQ Specific & BMQ General

1. Abstract

The Beliefs about Medicine Questionnaire (BMQ) is an internationally recognized, psychometrically validated self-report instrument designed to assess cognitive representations, beliefs, and attitudes regarding pharmacological treatments. Developed by Professor Robert Horne and colleagues in 1999, the instrument comprises two conceptually distinct yet complementary parts: the BMQ-Specific and the BMQ-General. The BMQ-Specific consists of 10 items assessing personal beliefs concerning medication prescribed for a particular illness, structured into two 5-item subscales: Specific-Necessity (perceived personal need for treatment to maintain current and future health) and Specific-Concerns (worries about potential negative consequences, long-term dependence, disruptive effects, and adverse reactions). The BMQ-General consists of 8 items evaluating broader societal and personal beliefs about pharmaceuticals, structured into two 4-item subscales: General-Overuse (perceptions that healthcare providers excessively prescribe and over-rely on medicines) and General-Harm (beliefs that medications are inherently toxic, addictive, or unnatural poisons).

Both sections employ a 5-point Likert scale ranging from 1 (Strongly disagree) to 5 (Strongly agree). Psychometric investigations across diverse chronic illness populations—including asthma, diabetes, cardiovascular diseases, rheumatoid arthritis, HIV/AIDS, psychiatric conditions, and renal transplants—demonstrate robust internal consistency (Cronbach’s alpha typically ranging from 0.70 to 0.86 for Specific subscales and 0.65 to 0.81 for General subscales), stable test-retest reliability, and strong construct, criterion, and predictive validity. Crucially, the theoretical framework operationalized by the BMQ—specifically the Necessity-Concerns Framework (NCF)—allows researchers and clinicians to calculate a cost-benefit differential (Necessity minus Concerns), which reliably predicts medication adherence, treatment persistence, and therapeutic engagement. Translated and validated in dozens of languages globally, the BMQ remains an indispensable psychometric tool within health psychology, behavioral medicine, clinical pharmacy, and public health interventions.

2. Keywords

Beliefs about Medicine Questionnaire, BMQ, medication adherence, Necessity-Concerns Framework, health psychology, pharmacotherapy beliefs, self-regulation theory, chronic illness management, psychometrics, treatment compliance

3. Authors

The Beliefs about Medicine Questionnaire was originally conceptualized, operationalized, and psychometrically validated by:

• Robert Horne, PhD, MSc, BPharm, FRPharmS, FAcSS: Professor of Behavioural Medicine at the UCL School of Pharmacy, University College London (UCL), United Kingdom. Professor Horne is the founding Director of the Centre for Behavioural Medicine at UCL and an internationally renowned leader in patient adherence, illness perceptions, and personalized healthcare communication.
• John Weinman, PhD, FRCPE, FMedSci: Professor of Psychology as Applied to Medicine at the Institute of Psychiatry, Psychology & Neuroscience (IoPPN), King’s College London, United Kingdom. Professor Weinman is a pioneer in health psychology, specializing in illness cognitions, self-regulation, and behavioral interventions in chronic diseases.
• Maureen Hankins, PhD: Quantitative Psychometrician and Medical Statistician, previously affiliated with the Department of Psychology, University of Brighton, and King’s College London, United Kingdom.

Dutch Adaptation: Denise T. D. de Ridder, PhD (Utrecht University) and Nicolet C. M. Theunissen, PhD (TNO Prevention and Health, Leiden, Netherlands) spearheaded the formal psychometric validation and linguistic translation of the Dutch version (BMQ-NL) in 2003.

4. Purpose

The primary clinical and epidemiological objective of the Beliefs about Medicine Questionnaire is to systematically quantify implicit and explicit cognitive representations that patients harbor toward pharmacotherapy. Non-adherence to prescribed medication regimens represents one of the most pressing global public health challenges, contributing substantially to avoidable morbidity, excess mortality, disease progression, and escalating healthcare expenditures across acute and chronic health conditions. Historical models of adherence frequently attributed treatment default to cognitive deficits, forgetfulness, low health literacy, or socio-demographic deprivation; however, empirical evidence increasingly demonstrates that non-adherence is predominantly an intentional, self-regulatory decision driven by personal beliefs, perceived trade-offs, and lay models of illness and pharmacology.

The BMQ was engineered to provide a standardized, theoretically anchored framework capable of discerning why individuals with identical objective diagnostic severity and prescription regimens exhibit divergent medication-taking behaviors. Specifically, the scale serves three interrelated purposes:

• Clinical Diagnosis of Perceptual Barriers: By distinguishing between patient doubts regarding treatment necessity and explicit anxieties regarding side effects or pharmacological toxicity, the BMQ enables clinicians, pharmacists, and allied healthcare professionals to conduct granular behavioral diagnostics. Rather than dispensing generic compliance reminders, clinicians can tailor shared decision-making interventions to address specific misconceptions.
• Empirical Research and Behavioral Modeling: The instrument allows researchers to explore the structural pathways through which cognitive representations influence health behaviors. The BMQ facilitates testing complex interactions between illness perceptions (as measured by the Illness Perception Questionnaire), demographic variables, clinical outcomes, healthcare utilization, and quality of life.
• Intervention Evaluation: The BMQ serves as a sensitive, responsive outcome measure in randomized controlled trials evaluating digital health apps, cognitive behavioral therapies, motivational interviewing, and pharmacist-led educational interventions aimed at optimizing treatment engagement.

5. Psychological Construct

The Beliefs about Medicine Questionnaire operationalizes four fundamental psychological dimensions divided into two distinct structural modules: the BMQ-Specific and the BMQ-General. Each dimension reflects a distinct domain of cognitive appraisal regarding pharmacological interventions.

BMQ-Specific: Proximal Medication Appraisals

The BMQ-Specific evaluates cognitive evaluations directly tied to the medication regimen prescribed for a designated personal medical condition. It comprises two subscales:

• Specific-Necessity (5 items): Measures the respondent’s belief in the personal indispensability, therapeutic efficacy, and future health protective value of their prescribed pharmacological regimen. High scores indicate an internalized conviction that current wellness and future disease prevention depend strictly upon adhering to the prescribed regimen (e.g., “My health, at present, depends on my medicines”; “Without my medicines I would be very ill”). Low scores reflect skepticism regarding therapeutic utility, perceived symptom independence, or ambivalence about personal vulnerability.
• Specific-Concerns (5 items): Captures personal apprehensions, perceived burdens, and negative expectations directly elicited by taking the prescribed pharmacotherapy. This dimension covers worries about disruptive experiential interference with daily life (“My medicines disrupt my life”), fear of iatrogenic toxicity or unpredictable physiological damage (“I sometimes worry about long-term effects of my medicines”), emotional dread of chemical dependence (“I sometimes worry about becoming too dependent on my medicines”), and conceptual alienation or cognitive confusion regarding the medication’s nature (“My medicines are a mystery to me”).

BMQ-General: Distal Societal and Pharmacological Schemata

The BMQ-General captures broader cultural, societal, and philosophical belief structures regarding pharmaceuticals as a general class of therapeutic technology, independent of personal prescription status. It comprises two subscales:

• General-Overuse (4 items): Gauges the respondent’s perception of the medical establishment’s prescribing patterns, assessing beliefs that biomedical practitioners over-rely on pharmacological therapies at the expense of alternative therapeutic strategies (e.g., “Doctors use too many medicines”; “If doctors had more time with patients they would prescribe fewer medicines”; “Natural remedies are safer than medicines”). High scores signify widespread medical skepticism and distrust of institutional prescribing paradigms.
• General-Harm (4 items): Reflects deeply seated fears that pharmaceuticals are inherently toxic, dangerous, unnatural, or chemically detrimental substances that produce net negative physiological consequences (e.g., “Most medicines are poisons”; “Medicines do more harm than good”; “People who take medicines should stop their treatment for a while every now and then”). This subscale often acts as a cognitive antecedent to specific medication concerns.

6. Theoretical Framework

The Beliefs about Medicine Questionnaire is grounded in health psychology and behavioral medicine, drawing primarily upon Howard Leventhal’s Common-Sense Model of Self-Regulation (CSM) and Robert Horne’s Necessity-Concerns Framework (NCF).

The Common-Sense Model of Self-Regulation (CSM)

Leventhal’s CSM posits that individuals confronted with somatic symptoms or diagnostic labels construct subjective, parallel cognitive and emotional mental models of their illness. These cognitive representations encompass five core dimensions: identity (symptom labels), timeline (acute, cyclical, or chronic), cause (etiological attributions), consequences (anticipated life impact), and controllability/cure (efficacy of coping mechanisms). While the original CSM primarily focused on mental representations of illness, Horne extended this theoretical paradigm by arguing that individuals construct analogous cognitive representations of treatment. Treatment beliefs do not exist in isolation; they are dynamically calibrated against illness representations. For instance, an individual who conceptualizes hypertension as an intermittent, symptomatic condition is prone to doubt the continuous necessity of anti-hypertensive pharmacotherapy, viewing lifelong daily consumption as irrational.

The Necessity-Concerns Framework (NCF)

Horne integrated the CSM with social cognition and decision theory to formalize the Necessity-Concerns Framework. The NCF posits that medication adherence decisions reflect an internal, ongoing psychological cost-benefit balance. For any prescribed treatment, an individual weighs their subjective cognitive assessment of personal necessity against their subjective cognitive and emotional concerns:

When perceived necessity substantially outweighs concerns, adherence is robust. Conversely, when concerns equal or surpass perceived necessity, the cognitive cost-benefit ratio tilts, precipitating deliberate non-adherence, unauthorized dose reductions, or drug holiday behaviors. Research confirms that this model delineates four distinct behavioral quadrants:

• Accepting (High Necessity, Low Concerns): Optimal adherence trajectory; patients recognize therapeutic value and experience negligible anxiety.
• Ambivalent (High Necessity, High Concerns): High cognitive conflict; patients recognize the absolute need for medicine but experience severe distress over side effects or dependency, often resulting in erratic adherence.
• Indifferent (Low Necessity, Low Concerns): Low engagement; patients perceive little health vulnerability and harbor few active worries, leading to unintentional omission or passive discontinuation.
• Skeptical (Low Necessity, High Concerns): High resistance; patients reject treatment necessity while actively fearing medication toxicity, correlating with high rates of intentional non-adherence.

7. Validity

The psychometric validity of the BMQ has been extensively corroborated across diverse international cohorts, clinical settings, and health condition taxonomies.

Construct and Factorial Validity

In the seminal psychometric validation study by Horne, Weinman, and Hankins (1998, 1999), principal component analyses and structural equation modeling unequivocally substantiated the differentiation between the Specific and General modules, as well as the independent two-factor structures within each module. Correlational analyses between the BMQ-Specific and the Illness Perception Questionnaire (IPQ) demonstrated that Specific-Necessity beliefs correlated strongly and positively with perceptions of illness chronicity and perceived illness consequences, confirming that treatment necessity representations systematically reflect the severity and permanence of the underlying illness model.

Predictive and Criterion Validity

Numerous systematic reviews and meta-analyses have quantified the predictive validity of the BMQ. A landmark meta-analysis by Horne et al. (2013), incorporating 94 independent studies across 23 countries and over 25,000 patients with chronic diseases, demonstrated that:

• Higher Specific-Necessity scores were consistently associated with higher medication adherence (Pooled Odds Ratio [OR] = 1.742, 95% CI [1.569, 1.934], p < 0.0001).
• Higher Specific-Concerns scores were consistently associated with lower medication adherence (Pooled OR = 0.504, 95% CI [0.450, 0.564], p < 0.0001).
• The Necessity-Concerns Differential (Necessity score minus Concerns score) served as an even stronger predictive metric of electronically monitored adherence (MEMS track caps), pharmacy refill metrics (Proportion of Days Covered [PDC]), and biological biomarkers (e.g., viral suppression in HIV, HbA1c control in type 2 diabetes, and serum therapeutic drug monitoring in organ transplantation).

Convergent and Discriminant Validity

The BMQ-General subscales demonstrate robust convergent validity with generalized health locus of control scales, trust in physician scales, and holistic health worldview measures. Specifically, General-Harm and General-Overuse positively correlate with inclinations toward complementary and alternative medicine (CAM) and inversely correlate with trust in conventional institutional healthcare. Discriminant validity is evidenced by weak or non-significant correlations with unrelated demographic indices (such as income and formal educational attainment) and generic personality traits (e.g., extraversion, openness), establishing that the BMQ captures distinct cognitive schemas regarding pharmaceuticals rather than general psychological distress or socio-demographic status.

8. Reliability

The reliability of the Beliefs about Medicine Questionnaire has been demonstrated across multiple clinical and non-clinical cohorts.

Internal Consistency

Original validation data and subsequent cross-cultural adaptations report high internal consistency across all subscales, generally surpassing the standard psychometric benchmark of α ≥ 0.70:

• Specific-Necessity: Original cohort α = 0.74 to 0.86 across asthma, diabetes, renal, and cardiac groups. Cross-cultural adaptations (e.g., Dutch, German, Chinese, French, Spanish) consistently report α coefficients between 0.76 and 0.88.
• Specific-Concerns: Original cohort α = 0.65 to 0.80. Subsequent broad-scale chronic illness cohorts consistently document α coefficients between 0.72 and 0.84.
• General-Overuse: Original validation α = 0.69 to 0.78. Across diverse international samples, internal consistency typically ranges from 0.67 to 0.81.
• General-Harm: Original validation α = 0.60 to 0.71; updated versions and large sample translations report α coefficients routinely between 0.68 and 0.79.

Test-Retest Stability

Temporal stability assessments over intervals ranging from two weeks to six months confirm robust test-retest reliability in stable chronic illness populations, with Pearson’s r and Intraclass Correlation Coefficients (ICC) consistently exceeding 0.70 (e.g., Specific-Necessity r = 0.77; Specific-Concerns r = 0.76; General subscales r > 0.70). These metrics confirm that the BMQ captures enduring cognitive schemas, while remaining sensitive to targeted clinical interventions and structured patient educational programs.

9. Factor Analysis

The dimensionality of the BMQ has been scrutinized via both Exploratory Factor Analysis (EFA) and Confirmatory Factor Analysis (CFA) across numerous linguistic and cultural adaptations.

Original Exploratory Factor Analysis

In the scale development phase (Horne et al., 1999), separate principal component analyses with varimax rotation were performed on the Specific and General item pools:

• BMQ-Specific: Extraction yielded a clear two-factor solution accounting for approximately 44% to 51% of total variance across clinical samples. Factor 1 (Specific-Necessity) showed primary loadings ranging from 0.58 to 0.82 across items 1, 3, 4, 7, and 10, with negligible cross-loadings (< 0.25) on Concerns. Factor 2 (Specific-Concerns) showed primary loadings ranging from 0.52 to 0.79 across items 2, 5, 6, 8, and 9.
• BMQ-General: Extraction yielded an unambiguous two-factor solution accounting for approximately 49% to 55% of the total variance. Factor 1 (General-Overuse; items 11, 14, 17, 18) displayed loadings between 0.61 and 0.80. Factor 2 (General-Harm; items 12, 13, 15, 16) exhibited loadings between 0.56 and 0.83.

Confirmatory Factor Analysis (CFA) & Model Fit

Subsequent structural equation modeling across hundreds of validation cohorts has confirmed the adequacy of the correlated two-factor model for both BMQ-Specific and BMQ-General. Typical goodness-of-fit indices include:

• Comparative Fit Index (CFI): Routinely observed between 0.92 and 0.97, surpassing the standard 0.90 threshold.
• Tucker-Lewis Index (TLI): Typically reported between 0.90 and 0.96.
• Root Mean Square Error of Approximation (RMSEA): Consistently falls between 0.038 and 0.065 (with 90% confidence intervals well below 0.08), indicating good model fit.
• Standardized Root Mean Square Residual (SRMR): Values reliably below 0.055.

While some psychometric investigations have reported marginal cross-loadings for item 6 (“My medicines are a mystery to me”) or item 14 (“Natural remedies are safer than medicines”) across specific non-Western translations, measurement invariance testing (configural, metric, and scalar invariance) generally supports structural equivalence across genders, age groups, and chronic disease cohorts.

10. Instrument / Measurement Tool

• Instrument Name: Beliefs about Medicine Questionnaire (BMQ)
• Component Scales:
  • BMQ-Specific: Comprising Specific-Necessity (5 items) and Specific-Concerns (5 items).
  • BMQ-General: Comprising General-Overuse (4 items) and General-Harm (4 items).
• Total Item Count: 18 items (10 items in BMQ-Specific; 8 items in BMQ-General). Each module can be administered independently or together.
• Administration Format: Self-administered paper-and-pencil, digital/web-based questionnaire, or structured clinical interview.
• Target Population: Adults (≥18 years) and older adults managing chronic illness regimens; adaptable for pediatric parental proxy-reports.
• Average Completion Time: 5 to 10 minutes.
• Response Format: 5-point Likert scale:
  • 1 = Strongly disagree
  • 2 = Disagree
  • 3 = Uncertain
  • 4 = Agree
  • 5 = Strongly agree
• Scoring System & Rules:
  • Specific-Necessity Subscale: Sum or mean of items 1, 3, 4, 7, and 10. Raw score range: 5 to 25. Higher scores indicate stronger perceived necessity of prescribed medicine.
  • Specific-Concerns Subscale: Sum or mean of items 2, 5, 6, 8, and 9. Raw score range: 5 to 25. Higher scores indicate greater concern about potential adverse consequences, dependence, and disruption.
  • General-Overuse Subscale: Sum or mean of items 11, 14, 17, and 18. Raw score range: 4 to 20. Higher scores indicate stronger belief that medicines are overprescribed by physicians.
  • General-Harm Subscale: Sum or mean of items 12, 13, 15, and 16. Raw score range: 4 to 20. Higher scores indicate stronger belief that medicines are toxic, harmful, or intrinsically poisonous.
  • Necessity-Concerns Differential: Calculated as [Specific-Necessity Score − Specific-Concerns Score]. Range: −20 to +20. Positive scores reflect an adherence-favorable cognitive appraisal (necessity exceeds concerns), whereas negative scores indicate potential risk for intentional non-adherence (concerns exceed necessity).

11. Permissions & Fee and Test Year

The original Beliefs about Medicine Questionnaire was developed and published in 1999 by Professor Robert Horne and colleagues at King’s College London and the University of Brighton. The Dutch linguistic validation (BMQ-NL) was published in 2003 by Denise de Ridder and Nicolet Theunissen.

Copyright and Licensing Terms: The intellectual property rights to the BMQ are held by Professor Robert Horne and University College London (UCL) / UCL Business (UCLB). Conditions for use are categorized as follows:

• Academic and Non-Commercial Research: The instrument is broadly accessible for non-funded academic research, individual postgraduate study, and clinical practice, typically without licensing fees upon submission of an academic license agreement request. Researchers must register their study and obtain permission prior to implementation.
• Commercial Use and Funded Clinical Trials: For commercial entities, pharmaceutical manufacturers, funded research consortia, and contract research organizations (CROs), commercial licensing agreements and royalty fees administered through UCLB (UCL Business Ltd) apply.
• Official Inquiries: Permissions, language translation requests, and licensing terms can be obtained through the Centre for Behavioural Medicine, UCL School of Pharmacy, London, United Kingdom.

12. References

Below are primary academic references documenting the development, psychometric validation, and behavioral applications of the Beliefs about Medicine Questionnaire:

• de Ridder, D., & Theunissen, N. (2003). Beliefs about Medicine Questionnaire (BMQ): Nederlandse versie [Dutch version]. Utrecht University & TNO Prevention and Health.
• Foot, H., La Caze, A., Gujral, G., & Cottrell, N. (2016). The necessity–concerns framework predicts adherence to medication in multiple illness conditions: A meta-analysis. Patient Education and Counseling, 99(5), 706–717. https://doi.org/10.1016/j.pec.2015.11.004
• Horne, R., & Weinman, J. (1999). Patients’ beliefs about prescribed medicines and their role in adherence to treatment in chronic illness. Journal of Psychosomatic Research, 47(6), 555–567. https://doi.org/10.1016/S0022-3999(99)00057-4
• Horne, R., Weinman, J., & Hankins, M. (1999). The Beliefs about Medicines Questionnaire: The development and evaluation of a new method for assessing the cognitive representation of medication. Psychology & Health, 14(1), 1–24. https://doi.org/10.1080/08870449908407311
• Horne, R., Chapman, S. C. E., Parham, R., Freemantle, N., & Forbes, A. (2013). Understanding patients’ adherence-related beliefs about medicines prescribed for long-term conditions: A meta-analytic review of the Necessity-Concerns Framework. PLOS ONE, 8(12), e80633. https://doi.org/10.1371/journal.pone.0080633
• Leventhal, H., Phillips, L. A., & Burns, E. (2016). The Common-Sense Model of Self-Regulation (CSM): A dynamic framework for understanding illness self-management. Journal of Behavioral Medicine, 39(6), 935–946. https://doi.org/10.1007/s10865-016-9782-2
• Menckeberg, T. T., Bouvy, M. L., Bracke, M., Kaptein, A. A., Leufkens, H. G., Raaijmakers, J. A., & Horne, R. (2008). Beliefs about medicines predict refill adherence to inhaled corticosteroids. Journal of Psychosomatic Research, 64(1), 47–54. https://doi.org/10.1016/j.jpsychores.2007.07.016

13. Items of the Scale