A psychometric evaluation published in the Journal of Patient-Reported Outcomes provides new empirical support for using the Patient-Reported Outcomes Measurement Information System Medication Adherence Scale, known as the PMAS, among cancer patients prescribed oral anticancer therapy. Conducted as a secondary analysis of a longitudinal observational investigation, the study examined whether this standardized self-report instrument could reliably and validly capture medication adherence behaviors and beliefs in individuals receiving maintenance therapy for multiple myeloma. Oral anticancer medications have rapidly become a cornerstone of contemporary cancer treatment, transforming multiple myeloma from an acutely fatal hematologic condition into a manageable chronic blood cancer. However, because these oral regimens are self-administered at home rather than infused in clinical settings by healthcare professionals, optimal disease outcomes depend heavily on consistent long-term adherence. The research team evaluated the internal consistency reliability, structural dimensionality, convergent construct validity, and objective criterion validity of the scale against microelectronic monitoring data, offering crucial methodological insights into outpatient oncology care.
The transition toward oral anticancer regimens has reshaped modern oncology practice, creating an urgent clinical imperative for psychometrically robust monitoring tools. Between 2000 and 2022, the relative proportion of oral compared to intravenous cancer therapy approvals surged dramatically from 7 percent to 56 percent, with an estimated 40 to 50 percent of newly emerging oncologic therapeutics currently in development designed for oral administration. While life-long oral anticancer regimens such as immunomodulatory agents have significantly extended progression-free and overall survival in multiple myeloma, maintaining consistent therapeutic intake is essential. In hematologic malignancies, published clinical trial data indicate that consuming more than 90 to 95 percent of prescribed doses is a major predictor of favorable treatment response. Despite this necessity, non-adherence remains widespread across chronic conditions, where failure to take medications as directed can compromise efficacy, exacerbate symptom burden, heighten healthcare costs, and increase mortality. Patient-reported instruments offer unique diagnostic utility by identifying subjective reasons for missed doses, yet validated, open-access measures tailored to oncology populations have historically remained scarce.
To evaluate the performance of the nine-item PMAS questionnaire in this specialized patient population, investigators analyzed data collected from an Institutional Review Board-approved six-month observational study conducted across four outpatient oncology clinics within a Comprehensive Cancer Center in southwestern Pennsylvania. Clinical eligibility criteria required participants to be at least eighteen years of age, diagnosed with multiple myeloma, actively prescribed an oral anticancer medication as maintenance therapy, fluent in English, and cognitively competent to provide informed consent. Out of 125 patients approached for potential participation, eleven were deemed ineligible, and twenty-eight eligible patients declined enrollment, citing study burden or excessive personal stressors. Ultimately, eighty-four eligible individuals consented to participate. Following initial baseline withdrawals and subsequent study attrition due to changing clinical eligibility, participant preference, loss to follow-up, or death, seventy-four patients successfully completed follow-up survey assessments at both three and six months post-enrollment. A criterion validity subsample of sixty-seven participants supplied complete electronic event monitoring data alongside their completed questionnaires.
The patient cohort reflected characteristic demographic and clinical features of outpatient populations managing multiple myeloma. The overall sample had a mean age of 63.8 years, comprised 59.5 percent male participants, and predominantly identified as either non-Hispanic white (83.3 percent) or Black or African American (13.5 percent). Educational backgrounds, household income tiers, and neighborhood socioeconomic conditions represented a broad spectrum, with 97.2 percent reporting that their household income adequately met basic living needs. Participants presented with an average Eastern Cooperative Oncology Group performance status score of 1. At study enrollment, participants had lived with a multiple myeloma diagnosis for a median duration of 33.5 months, and the majority were actively receiving first-line or second-line oral therapies, specifically lenalidomide (66.2 percent) or pomalidomide (29.7 percent), which had been prescribed for a median treatment duration of 11.0 months. The nested subsample utilized for electronic criterion concordance testing displayed virtually identical clinical and demographic distributions, with only a slightly greater proportion receiving lenalidomide maintenance therapy.
The central questionnaire under investigation, the nine-item PMAS version 1.0, was originally developed using standard PROMIS qualitative and quantitative protocols and translated into multiple European languages. Structurally, the instrument captures two distinct conceptual domains through prorated subscale scores: Medication Beliefs and Knowledge, abbreviated as MBK, and Medication Taking Behaviors, abbreviated as MTB. The MBK subscale contains four declarative items rated across a five-point agreement scale ranging from strongly disagree to strongly agree, generating a subscale score from 4 to 20 without an explicit retrospective time limit. The MTB subscale contains five behavioral items evaluated on a five-point frequency scale ranging from never to always, yielding a score between 5 and 25 referencing a seven-day recall window. Summing both subscales produces a Total PMAS score spanning 9 to 45, where higher numerical values indicate superior adherence. Additionally, a standalone global adherence item assesses whether patients took medication as recommended over the preceding seven days, serving as an independent behavioral indicator.
To establish rigorous comparative criterion validity, researchers collected continuous, objective medication intake data using the AARDEX Medication Event Monitoring System, or MEMS. Each participant received medication bottles equipped with smart MEMS closure caps housing integrated microprocessors that recorded the exact date and timestamp whenever the container was opened and closed. These time-stamped access events served as presumptive evidence of dose administration. To safeguard data fidelity and reconcile potential discrepancies, participants maintained detailed daily paper diaries recording container refills, designated travel periods, physician-directed drug holidays, delayed doses, and incidental openings unrelated to pill ingestion. Research field notes and clinical medical records were cross-referenced with electronic logs to generate two standardized objective adherence metrics: the percentage of prescribed doses taken and the percentage of days with correct dosing intake. These objective indices were calculated across the initial ninety-day interval, the subsequent ninety-day interval, the comprehensive 180-day monitoring timeline, and a discrete seven-day window immediately preceding survey completion.
Beyond electronic pill-cap monitoring and self-reported adherence, the investigative protocol administered a comprehensive battery of validated instruments to characterize participants and test theoretical correlates of adherence. Sociodemographic attributes and perceived income adequacy were captured via structured self-reports, complemented by the University of Wisconsin-Madison Area Deprivation Index to quantify neighborhood-level socioeconomic distress. Clinical history and treatment regimens were extracted directly from electronic medical charts. Physical functional impairment was measured using the single-item Eastern Cooperative Oncology Group scale. Common cancer symptoms were evaluated through the ten-item revised Edmonton Symptom Assessment Scale, where participants scored their 24-hour symptom intensity on ten-point scales. Depressive symptoms over the preceding two weeks were tracked using the Patient Health Questionnaire-9, which incorporated proactive safety protocols for participants scoring ten or greater or exhibiting suicidal ideation. Cognitive concentration and memory abilities were measured through the European Organization for Research and Treatment of Cancer core questionnaire cognitive domain, while comorbid medical conditions were quantified using the self-reported Brief Charlson Comorbidity Index.
Statistical evaluation of the survey data demonstrated robust internal consistency reliability across repeated assessment intervals alongside pronounced ceiling effects. Total scale scores exhibited high stability over time, with 44 percent of participants experiencing no score change between the three-month and six-month visits, 31 percent showing slight score reductions, and 25 percent showing minor increases. Nunnally’s classical psychometric threshold of 0.70 was successfully met across all dimensions. For unweighted total scores, Cronbach’s alpha reached 0.82 with a 95 percent confidence interval of 0.75 to 0.87 at the three-month assessment, and 0.84 with a 95 percent confidence interval of 0.78 to 0.89 at six months. However, individual item distributions were strongly skewed toward optimal adherence, as more than three-quarters of respondents endorsed the highest possible rating on eight of the nine scale items. The lone exception was item four, which queries the belief that the medication is actively working, where fewer than 75 percent selected the maximum score, reflecting the clinical reality of an incurable, highly unpredictable malignancy.
Confirmatory factor analyses utilizing Bayesian estimation with empirically informed priors evaluated whether the empirical data conformed to a unidimensional structure or the hypothesized two-factor framework. The model fit statistics revealed striking structural divergence between the two constituent subscales. The single-factor model representing Total PMAS exhibited poor structural fit and substantially elevated Bayesian Information Criteria values across both measurement points. In sharp contrast, the Medication Beliefs and Knowledge subscale exhibited excellent structural fit, particularly at the six-month timepoint, where the Comparative Fit Index reached 1.00 and the Root Mean Squared Error of Approximation dropped to 0.00, although the authors cautioned that these extreme parameters might reflect a slightly overfitted solution. Conversely, the Medication Taking Behaviors subscale demonstrated consistently substandard fit indices. The investigators suggested that behavioral taking items function as formative rather than reflective indicators, aggregating heterogeneous daily routines that do not necessarily correlate with one another, unlike the cohesive cognitive construct underlying medication beliefs and disease knowledge.
Analysis of convergent construct validity supported the theoretical framework of the scale by replicating well-established clinical relationships. Because PMAS scores exhibited non-normal distributions and substantial ceiling clustering, researchers calculated non-parametric Spearman rank-order correlation coefficients. The observed associations aligned with initial empirical hypotheses: higher adherence scores demonstrated statistically significant positive correlations with advancing chronological age and superior self-reported cognitive functioning. Conversely, lower PMAS adherence scores demonstrated statistically significant negative associations with elevated physical symptom severity on the Edmonton scale and higher depressive symptom burdens on the Patient Health Questionnaire-9. At the same time, the empirical data revealed no significant statistical associations between adherence scale ratings and participant sex, racial background, annual household income, educational attainment, physical functional performance status, or general comorbid illness burden, indicating that the tool captures behavioral variance independent of broad sociodemographic categories.
Criterion concordance testing comparing self-reported survey scores with objective MEMS cap indices revealed statistically significant, weak-to-moderate positive correlations. Over the entire monitoring duration, median adherence captured by electronic event monitors reached 88.1 percent for prescribed doses taken and 86.6 percent for days with correct intake timing. Statistically significant positive Spearman rank-order correlations emerged primarily between objective electronic indices and PMAS metrics containing explicit behavioral content, including the Total score, the Medication Taking Behaviors subscale, and the single-item Global adherence question. Interestingly, evaluating the isolated seven-day objective monitoring window immediately preceding survey administration yielded no statistically significant correlation with self-reported seven-day recall items. The investigators concluded that self-reported questionnaires and microelectronic bottle monitors capture distinct facets of patient behavior rather than interchangeable data points, indicating that subjective surveys reflect broader patterns of beliefs and intentionality, whereas bottle cap processors record exact mechanical access events over prolonged intervals.
The study authors highlighted several methodological limitations that temper the immediate generalizability of these findings. First, the investigation relied on a relatively modest sample drawn from a single geographic healthcare system, accompanied by minor amounts of missing electronic monitoring data and constrained racial diversity. Second, pervasive ceiling effects in self-reported scores limited the statistical variance necessary to detect subtle behavioral declines or establish precise clinical cut-off scores that could flag non-adherent patients. Third, administering self-report surveys only at three and six months precluded granular tracking of transient behavioral lapses that were captured continuously by microprocessors. Furthermore, adherence dynamics in multiple myeloma may reflect unique psychological pressures, as patients managing an incurable malignancy may exhibit heightened vigilance when taking life-sustaining oral chemotherapy despite considerable financial and symptomatic challenges. The research team emphasized that future studies should employ larger, ethnically diverse cohorts, integrate objective neuropsychological assessments of cognitive function, and incorporate qualitative interviews to ensure that scale items capture the primary adherence barriers experienced by cancer patients.
Ultimately, this psychometric investigation establishes that the PROMIS Medication Adherence Scale is a feasible, reliable, and valid instrument for tracking oral anticancer adherence in routine oncology care and research. The scale provides clinicians with an open-access, inexpensive, and non-invasive mechanism to assess patient beliefs and identify practical behavioral barriers that might otherwise go undetected during brief clinical encounters. Given that modern oncology increasingly relies on prolonged oral regimens to sustain disease remission, identifying adherence challenges early is vital to preventing disease progression and avoidable mortality. However, because self-reported adherence exhibits known ceiling tendencies and only moderately correlates with microelectronic bottle monitoring, the study authors strongly advocate for multi-modal clinical monitoring programs. Integrating periodic self-reported assessments like the PMAS alongside objective verification methods, such as pharmacy refill ratios, electronic monitoring, and structured clinical reviews, offers the most effective strategy to optimize patient safety, support self-management, and maximize the therapeutic benefits of modern anticancer therapies.
Subject of Research: Psychometric evaluation of the PROMIS Medication Adherence Scale in multiple myeloma
Article Title: Psychometric evaluation of the PROMIS® Medication Adherence Scale among patients prescribed oral anticancer medication for multiple myeloma
Article References: Belcher, S. M., Scott, P., Sereika, S. M., Bender, C., Dunbar-Jacob, J., Rosenzweig, M. Q., Muluneh, B., Sabik, L. M., Copeland, V., Agha, M., Lustyik, A. A., He, E. S., Bozich, C., Booze, G., Duran Basso, P., & Peipert, J. D. (2026). Psychometric evaluation of the PROMIS® Medication Adherence Scale among patients prescribed oral anticancer medication for multiple myeloma. Journal of Patient-Reported Outcomes, 10(1), Article 172. https://doi.org/10.1186/s41687-026-01204-z
Image Credits: AI Generated
DOI: 10.1186/s41687-026-01204-z
Keywords: multiple myeloma, medication adherence, oral anticancer medications, PROMIS, psychometric evaluation, electronic event monitoring, MEMS caps, construct validity, confirmatory factor analysis, Cronbach alpha, lenalidomide, pomalidomide, patient-reported outcomes, symptom severity, cognitive function, oncology clinical research
Cite Scienmag News
Nathaniel Bowman. (November 20, 2026). Study Validates Self-Reported Adherence Scale for Multiple Myeloma Oral Anticancer Therapy. Scienmag. https://scienmag.com/study-validates-self-reported-adherence-scale-for-multiple-myeloma-oral-anticancer-therapy/
Nathaniel Bowman. "Study Validates Self-Reported Adherence Scale for Multiple Myeloma Oral Anticancer Therapy." Scienmag, 20 November 2026, https://scienmag.com/study-validates-self-reported-adherence-scale-for-multiple-myeloma-oral-anticancer-therapy/. Accessed 20 November 2026.
Nathaniel Bowman. "Study Validates Self-Reported Adherence Scale for Multiple Myeloma Oral Anticancer Therapy." Scienmag. November 20, 2026. https://scienmag.com/study-validates-self-reported-adherence-scale-for-multiple-myeloma-oral-anticancer-therapy/








