Epistemic conflict in neurodevelopmental assessment: a critical discourse analysis of second opinion requests

Background Adult ADHD and autism diagnostic services in the National Health Service (NHS) face severe capacity constraints. When assessments conclude that diagnostic criteria are not met, some individuals contest these decisions by requesting a formal second opinion. Objective This service evaluation examined how patients linguistically construct challenges to clinical diagnostic decisions in second opinion requests following negative ADHD or autism assessments. Methods We analysed 51 written second opinion requests submitted to an NHS adult neurodevelopmental diagnostic service using Fairclough’s critical discourse analysis framework, examining linguistic features, discursive practices, and social contexts. Demographic data were analysed using descriptive statistics; inferential tests were accompanied by effect sizes and an explicit power note. A trustworthiness framework comprising analytic memos, conservative coding, negative case analysis, peer debriefing within the interdisciplinary team, and an audit trail substituted for inter-rater reliability where a second coder was not feasible. Results The sample consisted of predominantly women (68.6%) with a mean age of 36.7 years. Four discursive strategies emerged: appropriation and recalibration of clinical taxonomy (98.0%), mobilization of biographical and pharmacological evidence (66.7%), hermeneutic reframing of clinical benevolence (98.0%), and temporal conflict of masking (58.8%). A fifth phenomenon, the competence paradox (86.3%), is reframed as a structural-discursive bind imposed by the genre rather than a patient strategy. The near-universality of each strategy individually, rather than any non-independent co-occurrence, characterized the rhetorical pattern. Inferential analyses on the self-referred subgroup ( n = 39) were underpowered: only effects of d > 1.0 were detectable. Reported effect sizes were small (Cramer’s V = 0.20; Cohen’s d = 0.12 for age by condition; d = 0.45 for age by gender), with confidence intervals consistent with both null and substantial effects. Conclusion Second opinion requests reveal three forms of epistemic injustice: testimonial (experiential knowledge discounted), hermeneutical (inadequate medical terminology), and a distinct procedural form (challenge genres systematically filter out those whose impairment makes genre-competent performance impossible and paradoxically render genre-competent performance as counter-evidence). These findings underscore the need for assessment pathways that prospectively capture masking, compensation, functional cost, and biographical impairment, supported by transparent diagnostic documentation, targeted assessor training, and accessible routes for reconsideration that reduce the competence paradox.