Challenges and contributing factors in glucocorticoid treatment adherence among patients with oral lichen planus: a cross-sectional study and management strategies

Abstract

Objective

To investigate the status of glucocorticoid treatment adherence in patients with oral lichen planus (OLP), identify its influencing factors using validated instruments, and propose a targeted management framework incorporating behavioral insights and digital health strategies.

Methods

A cross-sectional study was conducted among 116 OLP patients recruited from our hospital between January 2021 and December 2023. The study protocol was approved by the Hospital’s Ethics Committee (Approval No: 20201215), and written informed consent was obtained from all participants. Data were collected using a general information questionnaire, the Chinese version of the 8-item Morisky Medication Adherence Scale (MMAS-8), and a self-designed OLP knowledge questionnaire. Exploratory factor analysis was performed to validate the MMAS-8 structure. Multivariate logistic regression was used to identify independent predictors of poor adherence.

Results

The rate of poor adherence was 63.8% (74/116). Factor analysis confirmed a two-dimensional structure of non-adherence (unintentional and intentional). Multivariate analysis identified age ≥ 50 years, disease duration ≥ 3 years, recurrence frequency ≥ 3, glucocorticoid side effects, and rural residence as independent risk factors for poor adherence (all P < 0.05). Higher education level and household income served as protective factors.

Conclusion

Poor treatment adherence is prevalent among OLP patients and is influenced by multidimensional factors. The behavioral distinction between unintentional and intentional non-adherence provides a critical basis for precision management. We recommend an integrated strategy combining structured education, side-effect–optimized regimens, and digitally enabled follow-up to address adherence challenges and improve long-term outcomes.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12903-026-07899-2.

Introduction

Oral lichen planus (OLP) is a T-cell-mediated chronic inflammatory disease of the oral mucosa [1, 2], characterized by a prolonged course, recurrent attacks and a recognized risk of malignant transformation [3], posing ongoing challenges to long-term health management and patient quality of life [4–6]. Glucocorticoids, recommended as first-line therapy in consensus guidelines globally, rely on long-term and standardized medication adherence as the cornerstone of their efficacy [7, 8].

However, suboptimal treatment adherence is a pervasive “silent epidemic” in chronic disease management. For OLP, the combined effect of the disease’s characteristics and concerns regarding treatment side effects make adherence particularly problematic, constituting a critical bottleneck limiting therapeutic efficacy.

Current domestic research often remains at the level of simply describing adherence status, lacking systematic attribution of influencing factors. Few studies propose prospective, systematic management strategies based on empirical data and aligned with current research hotspots like “Internet + Healthcare” and intelligent chronic disease management. Therefore, this study aims to identify the real-world challenges of glucocorticoid treatment adherence in OLP patients through a rigorous cross-sectional survey, deeply explore its multidimensional causes, and subsequently build an integrated management strategy combining traditional interventions with digital technology.

Materials and methods

Study participants and ethics

A convenience sampling method was used to recruit OLP patients from the outpatient clinic of our Oral Mucosal Diseases Department between January 2021 and December 2023. Inclusion criteria: (1) The study primarily enrolled patients who were newly diagnosed with OLP at our department and had not previously received any topical or systemic glucocorticoid therapy for this condition. (Figs. 1 A, B), ; (2) Age ≥ 18 years; (3) Receiving glucocorticoid treatment (topical or systemic) for at least one month; (4) Clear consciousness and voluntary provision of informed consent. Exclusion criteria: (1) Comorbid other severe systemic autoimmune diseases (e.g., systemic lupus erythematosus, pemphigus vulgaris) or distinct oral mucosal diseases (e.g., oral lichenoid lesions, discoid lupus erythematosus, oral submucous fibrosis); (2) Absolute contraindications to glucocorticoid use; (3) Presence of cognitive impairment or mental illness. The diagnosis of OLP was established based on both accepted clinical features and characteristic histopathological findings, in accordance with the diagnostic criteria outlined by the World Health Organization (WHO) in its 2016 classification of oral potentially malignant disorders [9]. The study protocol was reviewed and approved by the Hospital Ethics Committee (Approval No: 20201215). Written informed consent was obtained from all individual participants prior to their enrollment.

Fig. 1.

Figure A shows a case of oral lichen planus in a patient. Figure B shows the biopsy pathological image

Sample size calculation

The sample size was calculated a priori using G*Power software (version 3.1) for a logistic regression analysis. Based on a pilot study and previous literature [10], we anticipated a medium effect size (Odds Ratio ≈ 2.5) for the key predictors. To achieve 80% power with a two-sided alpha of 0.05 and to account for up to 10 candidate predictor variables in the initial model, the calculation indicated a required sample size of approximately 150 participants. Accounting for an estimated 20% rate of invalid questionnaires, the target sample size was set at 180. However, due to the constrained recruitment period (3 years) and strict inclusion criteria, the final analyzable sample comprised 116 patients. Despite this, the Events Per Variable (EPV) ratio for our final model with 7 predictors and 74 events of poor adherence was 10.6, which meets and exceeds the conservative threshold of 10, ensuring the stability of the multivariate regression results .

Study instruments and quality control

General information questionnaire: self-designed, covering sociodemographic data, disease characteristics, and treatment details

To assess medication adherence, the Chinese version of the 8-item Morisky Medication Adherence Scale (MMAS-8) was selected for its brevity, validated reliability, and widespread use in evaluating adherence behaviors in chronic diseases. The scale has been translated and validated in Chinese populations, demonstrating good psychometric properties [11, 12, 20]: This scale consists of 8 items with a total score of 8. Scoring criteria: A score ≥ 6 was defined as “good adherence,” and a score < 6 as “poor adherence.” Widely used in chronic disease populations, the Chinese version has a Cronbach’s α of 0.83; in this study, it was 0.80, indicating good reliability. The corrected item-total correlations for all items ranged from 0.35 to 0.65, all exceeding the recommended threshold of 0.30, which further supports the homogeneity of the scale.

An exploratory factor analysis was performed to assess the structural validity of the Chinese MMAS-8 in our cohort. The Kaiser-Meyer-Olkin measure was 0.81, and Bartlett’s test of sphericity was significant (χ² = 352.18, p < 0.001), confirming the suitability of the data for factor analysis. Principal component analysis with varimax rotation, extracting factors with eigenvalues greater than 1, yielded a two-factor solution. These two factors collectively accounted for 58.4% of the total variance, demonstrating good structural validity of the scale in this study. Table 1.

Table 1.

Exploratory factor analysis loadings for the Chinese version of the 8-item Morisky Medication Adherence Scale (MMAS-8) (n = 116)

MMAS-8 Item	Factor 1 Loading	Factor 2 Loading	Communalities
1. Do you sometimes forget to take your medicine?	0.824	0.102	0.691
2. People sometimes miss taking their medications for reasons other than forgetting. Thinking over the past two weeks, were there any days when you did not take your medicine?	0.796	0.158	0.659
3. Have you ever cut back or stopped taking your medicine without telling your doctor because you felt worse when you took it?	0.221	0.768	0.639
4. When you travel or leave home, do you sometimes forget to bring along your medicine?	0.745	0.189	0.591
5. Did you take all your medicine yesterday?	0.682	0.238	0.524
6. When you feel like your symptoms are under control, do you sometimes stop taking your medicine?	0.304	0.715	0.602
7. Taking medicine every day is a real inconvenience for some people. Do you ever feel hassled about sticking to your treatment plan?	0.156	0.841	0.729
8. How often do you have difficulty remembering to take all your medicine?	0.809	0.225	0.707

Extraction Method: Principal Component Analysis. Rotation Method: Varimax with Kaiser Normalization. a. Rotation converged in 3 iterations. (Note: Factor loadings greater than 0.40 are considered significant and are in bold.)

OLP Knowledge Questionnaire (OLP-KQ)

A 15-item instrument was self-designed for this study to assess patient knowledge across key domains: disease nature, treatment goals, medication importance, side-effect management, and the significance of regular review [13]. To ensure content validity, the initial item pool was reviewed by a panel of five experts (three oral medicine specialists, two clinical methodologists), yielding a scale-level Content Validity Index (CVI) of 0.91. In the present sample, the questionnaire demonstrated acceptable internal consistency (Cronbach‘s α = 0.78). As a preliminary exploration of its relationship with behavior, the total knowledge score showed a weak but significant positive correlation with the MMAS-8 adherence score (r = 0.24, p = 0.01), offering initial evidence of criterion validity. Full construct validation and test-retest reliability assessment are recommended for future use. To ensure data quality, the questionnaire was administered face-to-face by uniformly trained investigators during patient follow-up visits.

Procedure of data collection

Data collection was conducted in a standardized manner by two trained research assistants. During the patient’s scheduled follow-up visit, the study was introduced, and written informed consent was obtained. Participants then completed a paper-based questionnaire packet in a private consultation room. The packet consisted of three parts in fixed order: (1) the General Information Form, (2) the Chinese MMAS-8, and (3) the OLP Knowledge Questionnaire. For patients with visual or literacy difficulties, the research assistant read each item aloud verbatim and recorded the responses without interpretation. Each questionnaire was checked for completeness on-site. The entire process required approximately 15-20 minutes per participant.

Statistical analysis

Statistical analysis was performed using SPSS software (version 25.0). Descriptive statistics were presented as means ± standard deviations or frequencies (percentages). To identify independent factors associated with poor treatment adherence, a binary logistic regression analysis was conducted. The dependent variable was adherence status, defined as ‘poor adherence’ (MMAS-8 score < 6) versus ‘good adherence’ (MMAS-8 score ≥ 6). Prior to finalizing the multivariate logistic regression model, multicollinearity among all independent variables was assessed using the variance inflation factor (VIF). A VIF value of less than 5 was considered indicative of no substantial multicollinearity. Variables were selected for inclusion in the multivariate logistic regression model using a two-step approach: (1) all variables yielding a p-value < 0.1 in the univariate analyses were initially considered; (2) additionally, variables deemed clinically relevant or supported by prior literature (e.g., gender, clinical type of OLP), regardless of their univariate p-value, were also included for a comprehensive assessment. The final model was constructed using a forward stepwise (LR) method. Results were expressed as odds ratios (ORs) with 95% confidence intervals (CIs), and a p-value < 0.05 was considered statistically significant.

The full versions of the Chinese MMAS-8 and the OLP Knowledge Questionnaire are provided in Supplementary File S1 and S2, respectively.

Results

Participant characteristics and adherence status

The clinical profiles of the patients are summarized in Table 2. The most common clinical presentation was the reticular type (43.1%), followed by the erosive type (38.8%). A majority of patients (65.5%) had lesions involving multiple oral sites, with the buccal mucosa being the most frequently affected location (87.9%).”

Table 2.

Clinical characteristics of the study participants (n = 116)

Clinical Characteristic	Category	n	%
Clinical Type	Reticular	50	43.1
	Erosive	45	38.8
	Plaque	21	18.1
Primary Location	Buccal Mucosa	102	87.9
	Tongue	35	30.2
	Gingiva	28	24.1
	Palate	15	12.9
Number of Sites	Single	40	34.5
	Multiple (≥ 2)	76	65.5
Disease Severity	Mild	48	41.4
	Moderate	52	44.8
	Severe	16	13.8

The sum of the location percentages may exceed 100%, as patients often have multiple affected areas

A total of 116 OLP patients were included, comprising 49 males (42.2%) and 67 females (57.8%), with a mean age of 51.8 ± 11.2 years. MMAS-8 assessment revealed a mean total score of 4.7 ± 1.6. According to the standard cutoff, 42 patients (36.2%) were classified as having good adherence (MMAS-8 score ≥ 6), while 74 patients (63.8%) had poor adherence (MMAS-8 score < 6). The detailed score distribution is presented in Table 3.

Table 3.

Distribution of Morisky medication adherence Scale-8 (MMAS-8) scores among OLP patients (n = 116)

MMAS-8 Score	Adherence Level	Frequency (n)	Percentage (%)	Cumulative Percentage (%)
8	High Adherence	10	8.6	8.6
7		15	12.9	21.5
6		17	14.7	36.2
5	Medium / Low Adherence	22	19.0	55.2
4		19	16.4	71.6
3		18	15.5	87.1
2		10	8.6	95.7
1		3	2.6	98.3
0		2	1.7	100.0

Among the 116 patients, 34 (29.3%) reported or exhibited glucocorticoid-related side effects. Of these, local side effects were predominant (n = 28), with oral candidiasis being the most common. A smaller subset reported systemic effects (n = 6).

Univariate analysis of factors influencing adherence

Univariate analysis showed that age, disease duration, number of disease recurrences, presence of glucocorticoid side effects, education level, family residence, and monthly household income were significantly different between the good and poor adherence groups (P < 0.05). Details are shown in Table 4.

Table 4.

Univariate analysis of factors associated with glucocorticoid treatment adherence in OLP patients [n(%)]

Factor	Category	Total (n = 116)	Good Adherence (n = 42)	Poor Adherence (n = 74)	χ²/t value	P value
Age (years)	< 50	53	25(47.2)	28(52.8)	5.624	0.018
	≥ 50	63	17(27.0)	46(73.0)
Disease Duration (years)	< 3	68	32(47.1)	36(52.9)	9.362	0.002
	≥ 3	48	10(20.8)	38(79.2)
Number of Recurrences	< 3	78	35(44.9)	43(55.1)	8.957	0.003
	≥ 3	38	7(18.4)	31(81.6)
Presence of Side Effects	No	82	36(43.9)	46(56.1)	10.235	0.001
	Yes	34	6(17.6)	28(82.4)
Education Level	High school or below	60	17(28.3)	43(71.7)	6.215	0.013
	College or above	56	25(44.6)	31(55.4)
Residence	Urban	72	32(44.4)	40(55.6)	7.103	0.008
	Rural	44	10(22.7)	34(77.3)
Monthly Household Income (CNY)	< 5000	50	13(26.0)	37(74.0)	7.785	0.005
	≥ 5000	66	29(43.9)	37(56.1)

Multivariate logistic regression analysis of factors influencing adherence

Diagnostic tests for multicollinearity using the variance inflation factor (VIF) showed all values were below 2.0, well under the conservative threshold of 5, indicating that multicollinearity was not a significant concern in our final model.

Multivariate logistic regression analysis, with adherence status as the dependent variable, confirmed that age ≥ 50 years, disease duration ≥ 3 years, number of recurrences ≥ 3, presence of glucocorticoid side effects, and rural residence were independent risk factors for poor adherence, while higher education level and higher monthly household income were protective factors [13]. Table 5. A comparison of the univariate (Table 4) and multivariate (Table 5) results shows that the strength and significance of associations for the identified independent factors remained robust after adjustment for other covariates. This consistency underscores their stable contribution to predicting poor adherence.

Table 5.

Multivariate logistic regression analysis of factors influencing treatment adherence in OLP patients

Factor	β	SE	Wald χ²	P value	OR	95% CI
Age ≥ 50 years	0.821	0.351	5.471	0.019	2.273	1.143–4.523
Disease Duration ≥ 3 years	0.935	0.368	6.455	0.011	2.547	1.238–5.241
Recurrences ≥ 3	1.105	0.398	7.703	0.006	3.019	1.384–6.584
Presence of Side Effects	1.254	0.425	8.705	0.003	3.504	1.523–8.063
Higher Education Level (Ref: Low)	-0.785	0.345	5.174	0.023	0.456	0.232–0.897
Rural Residence (Ref: Urban)	0.892	0.374	5.689	0.017	2.440	1.172–5.079
Higher Monthly Income (Ref: Low)	-0.811	0.358	5.130	0.024	0.444	0.220–0.897
Constant	-1.215	0.556	4.773	0.029	0.297

Discussion

Core contributions and synthesis of key findings

This study provides two primary contributions to the management of Oral Lichen Planus (OLP). First, it quantifies a critical clinical gap: a high rate of poor adherence (63.8%) to first-line glucocorticoid therapy. Second, and more importantly, it moves beyond simple description by applying a validated behavioral framework [6]. Exploratory factor analysis confirmed that non-adherence in OLP is not monolithic but consists of distinct “unintentional” and “intentional” pathways. This conceptual lens allows us to interpret the significant multifactorial associations from our multivariate model (Table 5) not merely as a list of risks, but as a map of potential mechanisms influencing patient behavior.

Interpreting key factors in clinical and behavioral context

A detailed interpretation of the independent factors from Table 5, viewed through the unintentional/intentional framework, offers nuanced clinical insights.

Side Effects & Intentional Non-Adherence: The strongest association (aOR = 3.50) was between glucocorticoid side effects and poor adherence. This finding directly explains a common clinical dilemma: the treatment’s primary toxicity becomes a major reason for its discontinuation. It underscores why merely prescribing medication is insufficient without proactive management of side effects (e.g., monitoring for oral candidiasis), aligning with strategies to use steroid-sparing agents like hydroxychloroquine [14, 15].

Clinical Course & Behavioral Pathways: Factors like frequent recurrences (aOR = 3.02) and longer disease duration (aOR = 2.55) are potent drivers of intentional non-adherence, likely stemming from therapeutic frustration and eroded belief in treatment efficacy. Conversely, older age (aOR = 2.27) is a classic risk factor for unintentional non-adherence due to polypharmacy and cognitive load.

Socioeconomic Barriers: The persistence of rural residence (aOR = 2.44) and lower income (aOR = 2.25) as independent factors highlights systemic inequities. These may impact both pathways—limiting access to follow-up (unintentional) and increasing the perceived burden and cost-benefit analysis of treatment (intentional).

Comparison with existing literature and clinical implications

Our findings both confirm and extend existing knowledge. The association of side effects and socioeconomic status with non-adherence is consistent with broader chronic disease literature [16, 17]. However, the validation of the two-factor behavioral structure in OLP is novel, providing a mechanism to explain why these factors matter. Compared to studies that report adherence rates alone, our analysis offers actionable etiology.

The clinical implications are direct and demand a stratified approach. For patients at risk of unintentional non-adherence (e.g., older adults), interventions should focus on reminder systems and simplifying regimens. For those prone to intentional non-adherence (e.g., patients experiencing side effects), the focus must shift to shared decision-making, effective patient education, and management of treatment-related concerns. A “one-size-fits-all” adherence intervention is thus inherently inadequate.

Consideration of key confounding variables

Our analysis intentionally treated “rural residence” and “experience of glucocorticoid side effects” as primary exposures of interest. However, we recognize their strong potential to act as confounders in the relationship between other sociodemographic factors and adherence. For instance, rural residence is often a proxy for multiple disadvantageous determinants, including limited access to specialist care, lower health literacy, and socioeconomic constraints. If not adequately controlled for, it could artificially amplify or obscure the apparent effect of income or education alone. Similarly, the experience of side effects may not only directly cause discontinuation but also alter a patient’s overall perception of treatment and their relationship with healthcare providers, potentially influencing their reporting of other behaviors or attitudes. In this study, we addressed these concerns by including both variables in our multivariate logistic regression model. This statistical adjustment allows us to estimate the independent association of other factors (e.g., age, income) after accounting for the influence of residence and side effects. Nonetheless, the possibility of residual confounding due to unmeasured aspects of these complex constructs (e.g., precise distance to clinic, severity of side effects) or other unknown factors cannot be ruled out.

Strengths, limitations, and future directions

The strengths of this study include its use of a validated, behaviorally informative scale (MMAS-8), robust multivariate analysis, and the novel integration of behavioral science into OLP management discourse. We transparently acknowledge its limitations: the cross-sectional design precludes causal inferences, and the single-center sample may affect generalizability. The sample size, while ultimately yielding a robust Events-Per-Variable (EPV) ratio of 10.6 in our final multivariate model—exceeding the conservative threshold for stability—was constrained by a defined recruitment period and the absence of a formal a priori power calculation. This may limit the generalizability of our findings and underscores the need for future prospective studies with pre-sampled size estimations to confirm our results.

This study has several limitations that also inform future research directions. First, the cross-sectional design precludes causal inferences regarding the identified associations. Second, the sample size was contingent on the prospective, consecutive enrollment in a single center over a specific period, which, while ensuring clinical relevance, may limit generalizability and the power to detect smaller effect sizes. Third, the use of convenience sampling at a single tertiary center may introduce selection bias, potentially limiting the external validity of our findings. Fourth, the OLP Knowledge Questionnaire, while developed with expert input for content validity, is a novel instrument lacking full psychometric validation. Its construct validity and test-retest reliability have not been established, which may affect the precision of the knowledge measurement. The reported correlations should therefore be interpreted as preliminary. This underscores the need for future studies to rigorously validate this or similar tools. To address these limitations, prospective longitudinal studies are needed to establish causality, and multicenter studies with larger cohorts would enhance external validity. Furthermore, incorporating objective adherence measures (e.g., pharmacy refill data) would strengthen the evidence. Ultimately, testing the proposed behaviorally-informed, digitally-enabled management framework in interventional trials is the logical next step to translate these findings into improved patient outcomes.

Conclusion

This study confirms that adherence to glucocorticoid treatment among patients with oral lichen planus is a significant challenge [18, 19], with multifaceted underlying causes. Based on robust attribution analysis and aligned with digital health trends, we have constructed a comprehensive management paradigm integrating structured education, personalized treatment, and intelligent follow-up. This paradigm not only provides concrete and actionable strategies for clinical practice but also lays the theoretical groundwork for future exploration of long-term, technology-enabled OLP management mechanisms, holding significant potential for clinical translation and broad application.

Abbreviations

OLP

Oral Lichen Planus

MMAS-8

8-item Morisky Medication Adherence Scale

OLP-KQ

OLP Knowledge Questionnaire

VIF

Variance Inflation Factor

EPV

Events Per Variable

OR

Odds Ratio

CI

Confidence Interval

Authors’ contributions

Conceptualization: BBG, YS，XQZ, MMM- Formal Analysis: BBG, XQZ, YXJ- Investigation: XQZ, YS，FQW，MMM, YNM.- Methodology: BBG, MMM, YXJ,- Project Administration: BBG, XQZ, FQW- Writing — Original Draft: BBG, XQZ- Writing — Review & Editing: BBG, XQZ,

Funding

No funding.

Data availability

The datasets used and/or analyzed in the current study areavailable from the corresponding author upon reasonable request.

Declarations

Ethics approval and consent to participate

This study was conducted in accordance with the principles of the Declaration of Helsinki. The study protocol was reviewed and approved by the Ethics Committee of The Affiliated Suqian Hospital of Xuzhou Medical University Approval No: 20201215. Written informed consent was obtained from all individual participants included in the study prior to data collection.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.