When “Choice” Is Not One Thing

Executive dysfunction, deliberate underuse, and undermedication in adults with ADHD

Evidence-graded refined deep-research report — September 2, 2026
Scope: Adults aged 18+, with emphasis on adults who knowingly or apparently take less ADHD medication than prescribed and on predominantly inattentive presentation (ADHD-PI) where the evidence permits.
Boundary: This is a descriptive and mechanistic review. It does not provide dosing guidance, titration schedules, or recommendations for any individual. Medication decisions belong with a prescriber.

1. Bottom line

1. “Deliberately taking less than prescribed” is not a single behavioral state. It can describe a prospective, values-based decision; a choice constrained by side effects, cost, scarcity, or regulatory friction; an implementation failure caused by prospective-memory and planning problems; or a retrospective explanation applied after a missed dose. Evidence: Strong for the conceptual distinction; Moderate for the component mechanisms; Weak for their population frequencies. [1–8]

2. The best quantitative adult dataset on reasons for stimulant nonadherence found that, among 395 adults classified as having low/medium adherence, forgetfulness was endorsed by 68.1%, long-term concerns by 52.4%, logistics by 47.3%, and medication beliefs by 44.8%. Respondents could endorse multiple domains. The overlap—not a single winning reason—is the central finding. Evidence: Moderate. [2]

3. Adult ADHD can impair the very processes required to carry out treatment: planning a future action, noticing the cue, initiating the action, managing refills, tracking whether a dose was already taken, and persisting through multistep administrative friction. In a laboratory study of 45 unmedicated adults with ADHD and 45 controls, the largest prospective-memory deficit was in planning, not simple recall. That supports a “treatment-task paradox,” but the study did not measure medication taking. Evidence: Moderate for the cognitive mechanism; Weak-to-Moderate for direct application to adherence. [7]

4. A 2026 cohort of 12,174,321 adults in seven countries, including 320,691 with ADHD, found that ADHD was associated with poorer adherence to an unrelated medication class: antihypertensive treatment. Poor-adherence odds were 1.45 (95% CI 1.26–1.67) at one year and 1.64 (1.34–2.00) at five years; within the ADHD group, ADHD-medication exposure was associated with lower poor-adherence odds, 0.66 (0.60–0.73) and 0.58 (0.46–0.72), respectively. This strongly supports a cross-medication implementation vulnerability, but it does not prove that ADHD medication causally fixes adherence: PDC measures possession, and healthcare engagement and other time-varying factors may confound the association. Evidence: Moderate–Strong for association; Weak for the specific mechanism. [34]

5. Genuine intentional underuse also exists. Adults report tailoring stimulant use to particular tasks, weekdays, social contexts, side-effect tolerability, authenticity concerns, stigma, cost, and perceived need. A 2025 qualitative systematic review synthesized 47 studies; a separate meta-ethnography synthesized 13 papers with 263 participants. These establish recurring motives but cannot estimate how common any motive is. Evidence: Moderate. [4–6]

6. “Choice” often remains real but bounded. Shortage, refill rules, pharmacy opacity, insurance, prescriber access, and the cognitive cost of repeatedly solving access problems narrow the feasible options. In a 20-participant shortage interview study, all reported pharmacy stock problems; 16 described calling multiple pharmacies, 10 restrictive refill policies, and 17 relapse of ADHD symptoms. Evidence: Moderate for structural constraints; Weak for the further claim that patients routinely reframe them as preference. [16]

7. Research methods are poorly suited to identify motive. Refill measures show possession, pill counts show missing quantity, electronic caps show container opening, self-report shows remembered explanations, and blood or urine tests show recent exposure. None establishes why a particular dose was or was not taken. Evidence: Strong. [1, 11, 14]

8. There is no validated adult syndrome, biomarker, performance test, or questionnaire that distinguishes under-titration from intentional underuse, executive failure, coverage mismatch, comorbidity, or structural scarcity. The most defensible research approach is triangulation across event timing, symptom and function measures, access history, and multiple adherence measures. Evidence: Strong. [1, 11, 14, 28–31]

9. Adult ADHD-PI is a major blind spot. The mechanisms most relevant to this review—planning, activation, time management, working memory, cognitive disengagement, and limited self-monitoring—map closely onto inattentive complaints, yet adherence and medication studies rarely stratify results by presentation. Evidence: Strong for the gap; Weak for any claim that ADHD-PI has a unique underuse phenotype. [7–9, 15, 29]

Evidence grades used in this report

• Strong: replicated meta-analytic evidence, large registry evidence with appropriate caveats, or a well-established measurement limitation.
• Moderate: consistent observational, qualitative, psychometric, or single-trial evidence.
• Weak: indirect, small, poorly replicated, or substantially extrapolated evidence.
• Contested: credible findings point in different directions.
• Anecdotal: patient-community or case material useful for phenomenology or belief mapping, not factual prevalence or causality.

These are review-level judgments, not a formal GRADE assessment.

2. The central correction: intentionality is event-level, layered, and time-dependent

The ordinary adherence vocabulary divides behavior into “intentional” and “unintentional” nonadherence. That split is administratively convenient but psychologically crude. An adult can have a general intention to follow a prescription, decide in advance to skip on a low-demand day, forget the next dose after a routine disruption, fail to refill because the task expanded into six pharmacy calls, and later describe the whole pattern as “I take it only when I need it.” Each statement can be partly true.

A five-layer analytic model proposed in this review

This model is evidence-informed but not a validated clinical instrument. It is designed to prevent a visible behavior—less medication consumed—from being mistaken for a single motive.

A sixth layer—feedback—then changes the next event. A missed dose may worsen organization, increase side effects on restarting, create surplus medication, reinforce doubts, or expose that the person functions differently than expected. The person’s next decision is therefore not independent of the prior miss.

Five forms of “taking less” that should not be collapsed

Key proposition: A patient can be choosing and impaired at the same time. “Choice” does not imply unconstrained opportunity, intact execution capacity, stable preferences, or accurate retrospective insight.

3. Definitions: the seven exposure states, revised around intentionality

Methodological flag: Most cited studies cannot distinguish all seven. Trials measure symptom change, claims measure dispensing, qualitative work measures reported experience, and adherence scales measure remembered reasons. None directly observes the full sequence from opportunity through intention, execution, and later explanation.

4. Executive dysfunction masquerading as choice

4.1 What the phrase means—and what it does not mean

“Executive dysfunction masquerading as choice” is not a diagnosis and should not be used to deny agency. It refers to situations in which a behavior looks voluntary from the outside—or is described as voluntary after the fact—but was materially shaped by impaired planning, cue detection, initiation, working memory, time management, monitoring, or persistence.

The phrase covers three different propositions:

1. Observed behavior is not motive. A skipped dose can look identical whether planned, forgotten, unaffordable, unavailable, or intentionally saved.
2. Self-description is evidence but not an infallible recording. Adults with ADHD can report their experience meaningfully, while still showing domain-specific limits in attention-related self-monitoring. [8, 9]
3. Agency can be bounded without disappearing. A person may genuinely decide not to complete a refill process after the process has become disproportionately difficult because of ADHD and structural friction.

The evidence is strongest for the first proposition, moderate for the second, and weaker for estimating how often the third changes a stated “preference.”

4.2 The treatment-task paradox

ADHD treatment adherence demands precisely the operations ADHD can impair:

• Prospective planning: forming a workable plan for a future action.
• Cue detection: noticing the time, context, or transition associated with the action.
• Goal maintenance: keeping the medication intention active while another task captures attention.
• Initiation: converting “I should” into physical action.
• Working-memory tracking: remembering whether the action has already occurred.
• Temporal organization: acting within a window rather than after it no longer feels useful.
• Multistep persistence: completing appointment, authorization, pharmacy, payment, and pickup steps.
• Error recovery: restarting the process after a missed call, stock failure, or rejected claim.

Fuermaier and colleagues compared 45 unmedicated adults with ADHD with 45 matched controls on a complex prospective-memory task. The largest impairment was in task planning; plan recall, self-initiation, and execution showed negligible-to-small group effects. [7] This does not prove that adult medication nonadherence is mainly a planning problem. It does show why “they remembered the prescription, therefore the miss was a choice” is too simple: remembering what to do differs from constructing and maintaining a plan that makes it happen.

4.3 Mechanism-by-mechanism analysis

4.4 The mixed-intentionality sequence

A single dose event can move through several states:

General intention to adhere → routine disrupted → cue missed → dose remembered late → expected benefit window feels reduced → person deliberately decides not to take it → later reports “I chose to skip.”

The final omission is intentional, but the pathway was initiated by executive failure. Calling it wholly intentional loses the proximal cause; calling it wholly unintentional erases the actual late-stage decision.

Another common structure is:

Medication supply becomes uncertain → person reserves doses for high-demand days → selective use becomes habitual → weekends are reclassified as “days I do not need it” → home, relationship, driving, and administrative impairments receive less weight because they are less formally evaluated than work output.

Here, the behavior is strategic and deliberate, but the preference may be an adaptation to scarcity rather than the person’s first-choice regimen. Direct evidence for the full reclassification process is lacking; shortage and situational-use studies establish its components. [4–6, 16]

4.5 Why “not a priority” is an ambiguous answer

In the Spalding study, “I do not consider taking the medicine a high priority in my daily routine” was associated with a mean 3.34 missed days in the prior week among respondents who endorsed that item. [2] This statement can mean at least five different things:

• a stable value judgment that medication is not important;
• a situational decision that other goals matter more on some days;
• avoidance because adverse effects are immediate;
• a planning failure expressed in moral or priority language;
• adaptation to low supply or access burden.

The item does not distinguish them. Treating “priority” as pure preference is therefore an interpretive error.

4.6 Insight limits: real but not global

It is tempting to explain underuse by saying adults with ADHD “lack insight.” The evidence does not support that broad claim. Butzbach et al. found small deficits in awareness of attentional performance, not global deficits across executive function and memory; self-versus-informant cognitive ratings did not significantly differ. [8] Kooij et al. found that adults were often the most useful informants about their ADHD, while tending to underreport severity relative to close others. [9]

The defensible conclusion is narrower:

• self-report is indispensable;
• retrospective self-report can miss event-level causation;
• awareness may be weaker for attention and gradual functional drift than for salient adverse effects;
• collateral data can add information but is not automatically more accurate;
• no insight measure can determine that a specific omission was secretly unintentional.

Evidence: Moderate.

4.7 Evidence that implementation generalizes—and changes with scaffolding

Two studies move the evidence beyond retrospective explanations, while still falling short of directly observing motive.

First, the vulnerability generalizes beyond ADHD medication. Yao et al. analyzed electronic health data from seven countries for 12,174,321 adults beginning antihypertensive therapy, 320,691 (2.6%) of whom met the study definition of ADHD. ADHD was associated with a higher five-year discontinuation rate, HR 1.14 (95% CI 1.02–1.27), and higher odds of PDC below 80% at one year, OR 1.45 (1.26–1.67), and five years, OR 1.64 (1.34–2.00). Among adults with ADHD, ADHD-medication use was associated with lower odds of poor antihypertensive adherence at one year, OR 0.66 (0.60–0.73), and five years, OR 0.58 (0.46–0.72). [34]

This is important because beliefs about stimulant identity, authenticity, addiction, or immediate stimulant side effects cannot by themselves explain poorer adherence to antihypertensives. The study therefore supports a broader implementation vulnerability. It does not identify executive dysfunction as the cause: ADHD was defined from diagnosis or medication dispensing; PDC measures medication possession rather than ingestion; confounding by healthcare contact, comorbidity, socioeconomic factors, and treatment engagement remains; and the first-discontinuation association was not observed in the young-adult stratum. Evidence: Moderate–Strong for a cross-medication association; Weak for a specific causal pathway.

Second, refill behavior can change when the treatment task is scaffolded. In a matched, nonrandomized study of adults aged 18–55, an ADHD-focused SMS program was associated with timely stimulant refills in 68% of participants versus 34% under usual care, OR 4.04 (95% CI 2.49–6.56). [35] A fixed-preference account cannot fully explain a large response to added prompts and engagement. But the intervention was not a pure memory reminder: it may also have changed knowledge, motivation, perceived support, and access to care, and refill timing still does not prove ingestion. Evidence: Moderate that adherence behavior is modifiable by scaffolding; Weak that the effect isolates executive dysfunction.

Together, these studies make one inference more defensible: some medication underuse is an implementation problem responsive to context, not merely a stable preference. They still do not allow a researcher—or an observer—to assign that mechanism to an individual dose event without event-level evidence.

5. What adults deliberately taking less report—and how to interpret it

5.1 Direct quantitative evidence: reasons overlap

Spalding et al. surveyed 602 U.S. adults who self-reported an ADHD diagnosis and at least three months of oral psychostimulant treatment. The Medication Adherence Reasons Scale classified 395 as low/medium adherence if they endorsed any of 19 reasons and 207 as high adherence if they endorsed none. [2]

Among the 395 low/medium-adherence respondents:

The mean number of prescribed days taken in the prior week was 4.56 ± 2.07 in the low/medium group versus 6.86 ± 0.70 in the high-adherence group. Because the four domains overlapped, their percentages must not be summed. [2]

5.2 The 19 reported reasons, reclassified by proximal mechanism

The means below are the mean number of days missed among respondents who endorsed the reason; they are not prevalence estimates. The right-hand classification is an analytic interpretation made in this report, not the original scale’s validated factor structure.

This table shows why “deliberate underuse” is not separable from the broader adherence system. Several reasons that sound like beliefs (“priority,” “need,” “working”) can be downstream of poor treatment fit, insight limits, routine collapse, or scarcity. Several reasons that sound unintentional can culminate in an intentional decision once the usual time has passed.

5.3 Qualitative evidence: adults actively tailor exposure

The 2025 qualitative systematic thematic review by Johnson et al. included 47 studies published from 1994 through 2023. It identified four broad themes: initial motivation and diagnosis, perceived benefits, context of use, and barriers. Adults described tailoring medication to tasks or times; stigma, cost, side effects, dependency concerns, and desire for self-sufficiency shaped use. [4]

Reeves and Tickle’s 2025 meta-ethnography included 13 papers and 263 participants. Its higher-order themes concerned performance and social fit, the burden of ADHD versus stimulant side effects, social functioning, and difficulty accessing medication and specialist care. Access to a specialist willing to adjust medication and dosage appeared to influence adherence and other treatment experiences. [5]

These reviews establish that selective use is meaningful, recurrent, and socially situated. They do not establish:

• the proportion of all adults who use medication selectively;
• the proportion whose stated reason matches the proximal cause of each miss;
• whether selective use produces better or worse net outcomes for a given person;
• a separate pattern for ADHD-PI.

5.4 Weekday/weekend use: clear evidence of situational choice, narrow population

Gould and Doucette surveyed 53 undergraduates prescribed ADHD medication. Students chose medication more often on weekdays than weekends. Weekday adherence related to beliefs about academic and social benefits; weekend adherence related more to negative side effects. Belief that medication caused loss of an authentic self reduced adherence in both contexts. [6]

This is one of the clearest demonstrations that adults can make day-specific cost–benefit decisions. Its limits are equally important: the sample was small, young, and enrolled in college. The findings cannot be generalized to middle-aged adults, nonstudents, or ADHD-PI specifically.

5.5 Necessity-versus-concerns reasoning

Khan and Aslani conducted three focus groups with 20 adults in Sydney. Participants described benefits, fear of functioning without medication, and lack of alternatives on one side; and side effects, stigma, dependence concerns, cost, and forgetfulness on the other. Forgetfulness was not merely “poor memory”: participants discussed uncertainty about whether they had already taken medication and disruption when sleep or eating routines changed. [3]

The useful inference is that adherence emerges from an interaction:

Perceived necessity × anticipated benefit × feasible execution competing with side effects × identity/social cost × access cost × uncertainty.

No single term in that equation is fixed. A high-demand day raises perceived necessity; a bad night’s sleep may raise concern about adverse effects; a shortage raises the option value of saving a dose; a prior missed dose can worsen the ability to complete the next refill.

6. A taxonomy of deliberate underuse, centered on reasoning rather than morality

6.1 Evidence-grounded or internally coherent reasoning

6.2 Identity, values, and social reasoning

6.3 Epistemic beliefs: accurate, exaggerated, or unresolved

6.4 Executive dysfunction as a recursive cause

6.5 Structural friction reframed as preference

The evidence directly supports structural friction; it only indirectly supports preference reframing.

In the shortage interview study of 20 affected individuals: [16]

• all 20 reported pharmacies being out of stock;
• 16 reported calling multiple pharmacies;
• 9 traveled long distances;
• 6 received partial fulfillment;
• 10 described restrictive refill policies;
• 17 reported relapse of ADHD symptoms;
• 13 reported reduced occupational performance;
• 11 reported emotional instability/dysregulation;
• 8 reported relationship problems.

This study included four affected minors represented by parents and was recruited through ADHD forums, so it is not an adult population estimate. It nevertheless demonstrates an important mechanism: policies that require repeated planning, calls, travel, and narrow timing windows impose their greatest burden on people whose condition affects planning and persistence.

The stronger claim—“patients routinely come to prefer the regimen the system forces on them”—has not been measured directly. The evidence supports adaptation, rationing, and situational tailoring. Calling that “preference adaptation” is a reasonable inference, not an established prevalence finding.

7. The stated-versus-actual-reason gap

7.1 What “actual reason” would mean

For a particular omitted dose, at least four kinds of cause can coexist:

1. Distal reason: beliefs, identity, concern about harm, or general medication attitude.
2. Contextual condition: weekend, low-demand day, poor sleep, social setting, shortage, travel.
3. Proximal mechanism: missed cue, remembered late, no supply, adverse effect, explicit planned skip.
4. Retrospective narrative: how the person explains the event later.

Most studies ask about the first or fourth and infer the third. That inference may be correct, partly correct, or wrong.

7.2 Why self-report alone cannot close the gap

Self-report is the only direct route to values and subjective reasons, but it has predictable limitations:

• Recall compression: repeated events are summarized into a general rule.
• Telescoping: a recent or salient event stands in for the month.
• Social desirability: “I don’t need it” may feel more acceptable than “I forgot again,” while “I forgot” may feel more acceptable than “I disliked how I felt.”
• Moral vocabulary: executive difficulty can be labeled laziness, irresponsibility, or lack of priority.
• Outcome-informed reconstruction: feeling fine on an unusually low-demand day can retrospectively validate the skip.
• Poor proximal memory: the person may remember that a dose was missed but not the sequence that produced it.

These are established general measurement concerns and fit the ADHD evidence, but direct adult studies comparing contemporaneous reasons with later explanations are essentially absent. Evidence: Strong for the measurement problem; Weak for the frequency of any specific distortion.

7.3 What objective measures add—and still miss

A 62-person randomized crossover study comparing immediate- and extended-release mixed amphetamine salts illustrates measurement disagreement: self-report and pill count did not detect a formulation adherence difference, while electronic monitoring did. Forty-nine completed the study. [14] Even electronic monitoring recorded openings, not ingestion or motive.

7.4 What a better research design would require

This is a research agenda, not a clinical protocol. A credible study of “executive dysfunction masquerading as choice” would combine:

• a clear prescribed plan;
• pharmacy availability and access events;
• electronic or digital dose-event measurement;
• pre-event intention captured before or near the intended window;
• event-level reason captured immediately after a miss;
• later retrospective explanation to measure narrative drift;
• sleep, demand, side effects, mood, substance use, and context;
• symptom and functional measures across work, home, relationships, and driving;
• collateral input where appropriate;
• stratification by ADHD presentation, sex/gender, age, formulation, comorbidity, and shortage exposure.

No identified adult study combines all of these. Evidence: Strong for the design gap.

8. Subjective signs of partial exposure, with special attention to executive failure

Residual symptoms do not reveal why exposure is insufficient. They can result from under-titration, missed doses, selective coverage, duration mismatch, comorbidity, sleep, cognitive disengagement syndrome, or an outcome that is not fully medication-responsive.

Wearing off, rebound, and a missed or skipped dose

These are often conflated:

• Wearing off: benefit fades toward baseline as exposure declines.
• Rebound: a transient worsening beyond baseline after exposure falls.
• Missed/late dose: the expected exposure was not established or began later.
• Afternoon crash: a phenomenological label that can include wearing off, rebound, hunger, sleep debt, accumulated effort, or mood effects.

A person who selectively takes less may describe all four as “the dose is not enough,” while a person with adequate peak response but short coverage may start rationing because only certain hours feel worth covering. Time course helps distinguish mechanisms, but no validated adult algorithm converts symptom shape into a dosing conclusion. Evidence: Moderate.

9. Objective signs and instruments: what detects residual burden, not motive

The core measurement conclusion is negative but important: objective tests can document residual symptoms, function, timing, or exposure opportunity. They cannot tell whether an adult chose underexposure freely, chose it under constraint, forgot, remembered late, or reconstructed a reason later.

10. Differential guide: what can mimic “I am deliberately taking less because this dose is enough”

11. Consequences and asymmetries

11.1 Partial versus full response

Short-term adult trials show robust average improvement in core symptoms, but broader functioning and quality of life improve less and often remain abnormal. A 2025 quality-of-life meta-analysis of 17 RCTs and 5,388 participants found Hedges g values of approximately 0.51 for amphetamines, 0.38 for methylphenidate, and 0.30 for atomoxetine. [22] In a college crossover study, 86.4% showed symptom reduction with lisdexamfetamine, yet large differences from controls remained in symptoms and executive functioning. [33]

This matters for intentional underuse because a person may accurately perceive only partial value. “Medication helps” and “the benefit is worth taking every prescribed day” are different judgments.

11.2 Adherence and work outcomes

In the Spalding cross-sectional survey, low/medium adherence was associated with greater absenteeism and indirect costs after covariate adjustment. [2] Causality cannot be inferred: more severe ADHD may cause both lower adherence and worse work outcomes, and the study was industry funded. The result nevertheless aligns with the recursive hypothesis: poorer symptom control and poorer implementation can reinforce one another.

In a 13-week randomized trial dataset of 276 adults, lower capsule-count adherence predicted less improvement on the CAARS observer-rated score. [11] Trial adherence was high overall and capsule counts cannot verify ingestion or motive.

The 2026 multinational antihypertensive cohort is the strongest evidence that adherence vulnerability is not confined to attitudes toward stimulants: adults with ADHD had higher odds of poor adherence to antihypertensives, while ADHD-medication exposure within the ADHD group correlated with better antihypertensive adherence. [34] This remains observational and cannot establish that ADHD medication caused the difference. Separately, a matched SMS study found timely stimulant refills in 68% versus 34% under usual care, showing that refill behavior can be highly context-sensitive; the intervention was nonrandomized and may have changed several mechanisms at once. [35]

In the n=96 naturalistic follow-up, roughly half of those using psychostimulants at specialty-clinic discharge still used any psychostimulant three years later; among persisters, adherence was relatively good. Nonadherence was associated with worse general functioning, mood, and sleep. [15] Direction of causality remains uncertain.

11.3 Selective coverage leaves asymmetric life domains

Adult ADHD medication research and patient narratives often privilege work or school because those outcomes are visible and measured. Yet medication-period registry associations involve domains that occur outside work hours:

• In a U.S. claims cohort of 2,319,450 adults, motor-vehicle crash odds were lower in medicated months: OR 0.62 in men and 0.58 in women. [18]
• In a Swedish cohort of 25,656 people with ADHD, criminal-conviction hazard was lower during medication periods: HR 0.68 in men and 0.59 in women. [19]
• In Swedish registry analyses, stimulant treatment was associated with lower later substance-related events, with important observational caveats. [19]

These data do not estimate the risk of taking half a prescribed dose, skipping a Saturday, or leaving evenings uncovered. Dispensed-prescription periods are coarse and confounded. They do show why the assumption “only work needs treatment” cannot be treated as an evidence-based fact. Driving, home administration, finances, relationships, parenting, and safety continue after paid work ends.

11.4 Risks of taking more versus risks of taking less

The tradeoff is real and asymmetric:

• More exposure can increase adverse-effect burden and physiological risk for some people. [20]
• Less exposure can return symptoms and impair function in exposure-sensitive domains. [2, 11, 15–19]
• The marginal benefit and marginal harm are individual and time-dependent.
• Population evidence cannot translate this tradeoff into an individual dose decision.

The evidence therefore contradicts both moral extremes: deliberate reduction is not automatically irrational, and lower exposure is not automatically safer overall.

12. Adult ADHD-PI: what can and cannot be said

What is reasonably supported

• Planning, activation, working memory, time management, prospective memory, and cognitive disengagement are highly relevant to adults whose impairment is predominantly inattentive. [7–9, 29, 30]
• These same domains can interfere with medication implementation and refill completion.
• Core symptom scales may miss functional administration burden; executive and functional scales can reveal residual impairment.

What is not established

• That ADHD-PI adults deliberately underuse more often than combined-presentation adults.
• That a specific symptom cluster identifies underdosing in ADHD-PI.
• That hyperactive/impulsive symptoms respond at a lower exposure than inattentive symptoms in adults.
• That planning failures, forgetfulness, or preference reframing are uniquely severe in ADHD-PI.
• That strategic workday-only use has different consequences by presentation.

One small naturalistic study of 96 adults associated the combined subtype, not inattentive subtype, with nonadherence among persisters. [15] That isolated finding argues against assuming an obvious PI-specific adherence disadvantage. Presentation classifications also change over time and are often not reported in modern treatment studies.

Overall evidence: Strong for the gap; Weak for PI-specific conclusions.

13. What the evidence does not know

14. Research conclusions

The refined evidence supports a more precise description than “the patient chooses not to take the full prescription.” For research purposes, the minimum defensible claim is:

The observed exposure was lower than prescribed, and the patient reports one or more reasons. The degree to which the event reflected prospective preference, constrained choice, executive implementation failure, or retrospective reconstruction is usually unresolved.

The strongest adult evidence indicates that reasons overlap, selective use is real, executive and routine failures are common, structural barriers materially shape the available choices, and adherence vulnerability extends to at least one unrelated chronic-medication class. Refill behavior also changes under structured external support. [34, 35] The weakest link is the one most often presumed: that a retrospective statement cleanly identifies the proximal cause of every missed or reduced dose.

The central asymmetry is methodological. Researchers can count prescriptions, fills, openings, pills, symptoms, and work absences much more easily than they can observe intention unfolding over time. Until adult studies capture opportunity, pre-event intention, implementation, and later narrative separately, “executive dysfunction masquerading as choice” should be treated as a serious, evidence-informed hypothesis—not as a hidden motive that can be assigned to any individual.

15. Source list

The list prioritizes sources most load-bearing for this focused review. Sample sizes refer to analyzed or enrolled samples as reported by the source. “Cannot distinguish states” means the design cannot reliably separate under-titration, missed doses, deliberate restriction, coverage mismatch, pharmacokinetic variation, and structural access.

16. Relationship to the original report

This refinement preserves the original report’s broader conclusions on partial response, objective measurement, differential diagnosis, adverse-effect and structural tradeoffs, registry outcomes, tolerance, pharmacokinetics, hormonal evidence, and the severe lack of adult ADHD-PI-specific data. It changes the organizing center from “all forms of undermedication” to the narrower question of how prospective choice, executive implementation, structural constraint, and retrospective explanation interact when an adult takes less than prescribed.

The most important added elements are:

• the five-layer intention-to-execution model;
• event-level rather than person-level intentionality;
• the reclassification of the 19 MAR-Scale reasons by proximal mechanism;
• direct integration of prospective-memory and metacognition evidence;
• explicit analysis of mixed-intentionality sequences;
• a measurement matrix explaining why no current data source can verify motive;
• a sharper distinction between proven structural adaptation and unproven preference reframing;
• cross-medication evidence that ADHD-associated adherence vulnerability extends beyond stimulant-specific beliefs;
• evidence that refill behavior is modifiable under structured external scaffolding, without overclaiming the mechanism.