[{"content":"I\u0026rsquo;m a neurologist and data scientist working at the intersection of clinical neurology and quantitative research, generating real-world evidence (RWE) on amyotrophic lateral sclerosis from disease registries and routine-care real-world data (RWD). My work spans epidemiology, biostatistics, and disease modelling — longitudinal analysis of disease progression, survival analysis, and the methodological discipline needed to draw regulatory-grade conclusions from messy observational data. I use R and Python to build reproducible analyses answering clinically relevant questions.\nYou can browse my publications, recent talks, the software I maintain, and occasional writing on ALS, data science, and statistics.\n","date":"26 May 2026","externalUrl":null,"permalink":"/","section":"Alejandro Caravaca Puchades","summary":"","title":"Alejandro Caravaca Puchades","type":"page"},{"content":"Occasional notes on ALS research, registry methodology, and the data-science workflows that sit underneath them.\n","date":"26 May 2026","externalUrl":null,"permalink":"/blog/","section":"Blog","summary":"","title":"Blog","type":"blog"},{"content":"","date":"11 May 2026","externalUrl":null,"permalink":"/tags/als/","section":"Tags","summary":"","title":"ALS","type":"tags"},{"content":"","date":"11 May 2026","externalUrl":null,"permalink":"/tags/alsfrs-r/","section":"Tags","summary":"","title":"ALSFRS-R","type":"tags"},{"content":"","date":"11 May 2026","externalUrl":null,"permalink":"/tags/clinical-trials/","section":"Tags","summary":"","title":"Clinical Trials","type":"tags"},{"content":"","date":"11 May 2026","externalUrl":null,"permalink":"/series/measuring-disease-in-als-a-critical-appraisal-of-the-alsfrs-r/","section":"Series","summary":"","title":"Measuring Disease in ALS: A Critical Appraisal of the ALSFRS-R","type":"series"},{"content":"","date":"11 May 2026","externalUrl":null,"permalink":"/tags/outcome-measures/","section":"Tags","summary":"","title":"Outcome Measures","type":"tags"},{"content":"","date":"11 May 2026","externalUrl":null,"permalink":"/series/","section":"Series","summary":"","title":"Series","type":"series"},{"content":"","date":"11 May 2026","externalUrl":null,"permalink":"/tags/","section":"Tags","summary":"","title":"Tags","type":"tags"},{"content":"In Part I of this series, we established why functional composite scales occupy a privileged position among ALS outcome measures: they sample across biological subsystems simultaneously, they are feasible to administer at scale, and they change detectably within the timeframe of a clinical trial. The ALSFRS-R, introduced by Cedarbaum and colleagues in 1999, became the field\u0026rsquo;s answer to these requirements. Over the following twenty-five years it accumulated a track record that no alternative can currently match in terms of sheer breadth of evidence. Understanding both why it succeeded and where it is failing is essential context for anyone interpreting ALS trial results — or designing new ones.\nThe scale in brief # The ALSFRS-R consists of 12 items spanning four functional domains: bulbar function (speech, salivation, swallowing), fine motor function (handwriting, cutting food, dressing), gross motor function (turning in bed, walking, climbing stairs), and respiratory function (dyspnoea, orthopnoea, respiratory insufficiency). Each item is scored on a 0–4 ordinal scale, yielding a total score between 0 and 48. A score of 48 represents normal function; decline toward zero tracks progressive disability. In most natural history cohorts, patients lose approximately 1 point per month on average, though individual trajectories vary enormously.\nAdministration takes 5–10 minutes and can be completed by telephone, which proved critical for large multisite trials and, later, for decentralised study designs. The scale requires no specialist equipment and minimal rater training, making it deployable in clinical settings that would not support formal strength or respiratory testing.\nWhat the ALSFRS-R does well # Its primary virtue is sensitivity to change relative to survival endpoints. Functional decline precedes death by months to years, meaning the ALSFRS-R captures treatment-relevant signal in a timeframe that survival analysis simply cannot. The ALSFRS-R is among the most robust clinical predictors of survival in ALS, supporting its widespread use as a measure of disease trajectory.1\nReliability is reasonable when the scale is administered under standardised conditions. Interrater agreement is acceptable for most items, and the telephone administration format, while initially controversial, was shown to produce results comparable to in-person assessment.2 For an international disease with rare, geographically dispersed patients, this is not a minor advantage.\nThe scale\u0026rsquo;s ubiquity is itself a strength, though an uncomfortable one to acknowledge. Because the ALSFRS-R has been used in virtually every ALS clinical trial for over two decades, it enables cross-trial comparisons, natural history modelling, and retrospective analyses that would be impossible with a fragmented measurement landscape. Any new scale will take years to accumulate comparable evidence. The ALSFRS-R\u0026rsquo;s dominance is path-dependent, but the path dependency is real.\nWhere it breaks down # The list of limitations is long enough to constitute its own indictment. Some are well-recognised, and careful trials guard against them. But familiarity with a scale\u0026rsquo;s weak points cuts both ways: little in the methodology stands between a known weakness and a study designed to lean on it. Still others are probably underappreciated — and may be lurking in the bottomless pit that seems to swallow every promising trial on its way from phase II to phase III:\nInterrater variability is the limitation most often cited and, in practice, most often dismissed. Agreement under standardised conditions is genuinely good — but that is the wrong frame. What matters for trial inference is whether the residual noise from rater interpretation, site practices, and administration drift is small relative to the changes trials are powered to detect. Pooled multicentre data show that around 12% of patients exhibit implausible ≥5-point increases between consecutive visits, with prevalence by site ranging from 0% to 83% — variability of the same magnitude as plausible treatment effects.3\nTreatment-induced score changes unrelated to neurodegeneration are another well-recognised problem, but one whose significance for analysis deserves more attention than it gets. The scale registers clinical management as if it were disease biology, and it does so in both directions. Non-invasive ventilation (NIV) lowers the respiratory subscale — initiating nocturnal BiPAP progressively drops item R3 from 4 to 2, and tracheostomy drops it to 0 — not because the underlying disease has worsened, but because the scale penalizes the intervention itself; the direction of bias is therefore counterintuitive, in that care pathways that prolong survival into ventilator-dependent stages appear to accelerate ALSFRS-R decline. In the opposite direction, symptomatic interventions like anticholinergics or botulinum toxin for sialorrhoea mechanically improve the salivation item without altering disease biology. When such interventions are used differentially between arms — or simply initiated at different rates — the ALSFRS-R conflates genuine neuroprotection with symptomatic management, and in the NIV case can actively obscure it. With the small sample sizes typical of ALS trials, randomisation cannot be relied on to balance the arms on these factors; and because management practices vary between centres, the resulting imbalance is difficult to correct for after the fact.\nPoor standardisation across sites and trials remains a persistent issue despite published guidelines. The wording of individual items has varied across versions and translations; the threshold between adjacent scores is interpreted inconsistently; and the phone administration protocol is not uniformly applied. None of these individually are catastrophic, but collectively they add noise to an already noisy signal.\nFloor and ceiling effects compress detectable variation at the extremes of the scale: patients with early-stage or slowly progressive disease cluster near 48, while end-stage patients reach 0 on multiple items well before death, making the final months of disease nearly invisible to the scale. Most of this is neutralised in practice by the enrollment criteria of typical trials, which routinely exclude very slow progressors and very advanced patients \u0026mdash; but what remains of the floor effect interacts with informative censoring (below) and with trajectory non-linearity to distort treatment-effect estimates near the end of follow-up.\nInformative censoring due to death is a problem that is statistically serious and has traditionally been handled inadequately. Patients who die during a trial are censored from longitudinal ALSFRS-R analyses at their last observed score \u0026mdash; but death is not missing at random. It is the terminal expression of the very process the scale is trying to measure. Standard mixed-effects models that ignore this mechanism produce biased estimates of treatment effects, typically in the direction of underestimating decline and therefore underestimating treatment benefit. The appropriate analytical tools \u0026mdash; joint models for longitudinal and survival data, pattern mixture models \u0026mdash; exist and are used in some trials but have only recently become standard.4\nThe ordinality problem: why one point is not always one point # Two limitations deserve deeper treatment because they are methodologically distinct, frequently conflated, and together undermine the statistical foundations of most ALSFRS-R analyses.\nThe first is the ordinal structure of the scale. Each item is scored 0–4, but there is no reason to assume \u0026mdash; and considerable evidence to doubt \u0026mdash; that the functional difference between a score of 3 and 4 is equivalent to the difference between 1 and 2. The items are anchored to qualitative descriptions, not to physical measurements, and the spacing between anchors reflects clinical judgement rather than any psychometric calibration.\nThe figure below illustrates this concretely using the bulbar subdomain \u0026mdash; speech, salivation, and swallowing \u0026mdash; whose three items have strikingly different functional architectures. On the salivation item, a score of 2 represents barely perceptible excess secretion: the patient has lost very little functional capacity relative to the full range of the item. On the swallowing item, a score of 2 means the patient can no longer eat a normal diet and is approaching tube feeding \u0026mdash; a profound functional transition. Yet both scores contribute identically to the bulbar subdomain total. When all three items score 2, the subdomain total is 6 regardless of which items drove the decline and how much function was actually lost along the way.\nFigure 1. Illustrative representation of the unequal functional weight carried by each score step across the three bulbar subdomain items (speech, salivation, swallowing). Brick width is proportional to the functional change that step represents. The bottom panel shows that a subdomain score of 6 \u0026mdash; reached when all three items score 2 \u0026mdash; reflects very different amounts of functional loss depending on which items drove the decline. A salivation score of 2 represents minimal excess secretion; a swallowing score of 2 means dietary modification is already required. Both contribute identically to the subdomain total. The practical consequence is that treating the ALSFRS-R total score as a continuous, interval-level variable \u0026mdash; as virtually all linear mixed-effects models do \u0026mdash; is a modelling assumption that is unlikely to hold. A one-point difference at different points on the scale does not represent the same quantity of disease. The total score aggregates these inequalities across 12 items, compounding the distortion. When we fit a model and report a treatment effect of \u0026ldquo;0.3 points per month,\u0026rdquo; that number is an average over a heterogeneous set of functional transitions that may have very different clinical meanings.\nA one-point difference at different points on the scale does not represent the same quantity of disease. [\u0026hellip;] When we report a treatment effect of \u0026ldquo;0.3 points per month,\u0026rdquo; that number is an average over a heterogeneous set of functional transitions that may have very different clinical meanings.\nNon-linearity of the decline trajectory # The second problem is about time rather than scale. The conventional analysis \u0026mdash; a mixed-effects model with a linear term in time, the workhorse of ALSFRS-R trials \u0026mdash; has each patient declining at a roughly constant rate, so that a treatment effect collapses to a single difference in slopes: \u0026ldquo;points per month.\u0026rdquo; Nothing in the mixed-model framework forces this; the \u0026ldquo;linear\u0026rdquo; in linear mixed model refers to linearity in the parameters, not to a straight-line trajectory. It is simply the specification almost everyone reaches for.\nThe problem is that empirically the decline is not constant. Individual trajectories tend to be gently sigmoidal \u0026mdash; a shallower phase early on, a steeper middle, a flattening near the floor as items bottom out one by one \u0026mdash; and the instantaneous rate varies systematically between patients with baseline score, symptom duration at entry, and onset region. A cohort slope is therefore an average over curves of different shapes, not just different steepness.\nWhether that matters over the length of a real trial is a fair question, and the answer is: usually less than one might fear, which is exactly why the linear model has survived. A phase II often runs six months and a phase III a year or more; over a window that short, a single patient\u0026rsquo;s trajectory is frequently close enough to a straight line that the misspecification is second-order. The real exposure is cohort composition: when patients enter at a range of disease stages, the trial pools shallow early segments with steep mid-disease ones, and if the arms are not perfectly balanced on stage \u0026mdash; which, at ALS sample sizes, they often are not \u0026mdash; the estimated slope difference quietly absorbs some of that imbalance. Longer trials, lead-in periods, and designs that lean on a pre-randomisation slope (for enrichment or as a covariate) make the curvature harder to ignore, since that pre-slope is itself a straight line fitted to a curved segment.\nSo why not reach for splines, a quadratic term, or a proper nonlinear progression model? The obstacles are mostly not statistical:\nInterpretability. \u0026ldquo;The drug slowed decline by 0.3 points per month\u0026rdquo; is a sentence a clinician, a patient, and a regulator can all act on; a time-by-treatment interaction gives an effect that changes over follow-up, which is harder to headline. That part is fixable \u0026mdash; collapse it back to one number, e.g. the difference in area under the trajectory over the fixed window, arguably a fairer summary of cumulative benefit than a slope. What remains is that this number is a modelling choice rather than a reading off the data, and a less familiar quantity than the slope regulators have decades of experience interpreting \u0026mdash; so the resistance here is conventional, more than it is statistical.\nOverfitting. Flexibility is not free. With a few hundred patients per arm, a handful of visits each, and substantial dropout, there is often not enough information per patient to estimate curvature stably; knot count and placement turn into tuning knobs, and the conclusions can move with them. A misspecified-but-stable linear fit can beat a flexible-but-noisy one at the job that matters here \u0026mdash; detecting a between-arm difference.\nPre-specification and inertia. The accepted primary analysis is a pre-specified mixed model with a linear time term \u0026mdash; the \u0026ldquo;rate of decline\u0026rdquo; slope. A nonlinear primary endpoint is a harder sell at a regulatory meeting and has little precedent, so even the groups that do fit sigmoidal or disease-progression models tend to keep them in a secondary or descriptive role.\nThe honest summary is that linearity is a known approximation whose error is usually small over a fixed, enrichment-narrowed trial window \u0026mdash; but \u0026ldquo;usually small\u0026rdquo; is not \u0026ldquo;negligible,\u0026rdquo; the error is rarely quantified, and it compounds with informative censoring and floor effects precisely where the data are thinnest. The defensible responses are unglamorous: keep the follow-up window short enough that the approximation holds, stratify or adjust for baseline progression rate, and look at the residuals for curvature rather than assuming it away.\nSummary of limitations and trial consequences # Limitation Mechanism Trial consequence Mitigation Interrater variability Inconsistent threshold interpretation between raters and sites Inflated residual variance; reduced power Rater training, centralised scoring, telephone administration protocols, self-administered methodologies NIV / symptomatic treatment confounding Interventions influence subscale scores unrelated to neurodegeneration Spurious treatment effect if differential use between arms Pre-specified covariate adjustment; separate reporting of respiratory subscale Poor standardisation Item wording, scoring anchors, administration protocol vary across sites and versions Measurement error; limits cross-trial comparability Strict protocol adherence; single validated version per trial; self-explanatory scoring sheets Floor / ceiling effects Score compression at boundaries of the scale Reduced sensitivity at disease extremes; residual floor effect interacts with censoring Enriched enrollment; early trial entry criteria Informative censoring Death is non-random and correlated with outcome trajectory Biased estimates under standard mixed models; underestimation of decline Joint longitudinal-survival models; pattern mixture models \u0026mdash; rarely applied in practice Ordinal scale treated as continuous Unequal functional weight of steps across items and score range Loss of interpretability; potential bias in effect estimates None within the scale as currently used Non-linearity Sigmoidal individual trajectories; rate varies with disease stage Misspecified linear models; slope difference absorbs arm imbalance on stage Short follow-up windows; baseline progression-rate stratification; nonlinear modelling (rarely the primary endpoint) Multidimensionality Domains reflect distinct biological processes with potentially different treatment responses Domain-specific effects cancelled by aggregation; false negatives None within a total score framework \u0026mdash; requires modeling subdomain scores The last row of that table \u0026mdash; multidimensionality \u0026mdash; is listed here for completeness, but it warrants a dedicated discussion. It is, in my view, the most consequential and least adequately addressed limitation in the current trial literature, and it is the subject of Part III.\nKimura et al. (2006). Progression rate of ALSFRS-R at time of diagnosis predicts survival time in ALS.\u0026#160;\u0026#x21a9;\u0026#xfe0e;\nKaufmann et al. (2007). Excellent inter-rater, intra-rater, and telephone-administered reliability of the ALSFRS-R in a multicenter clinical trial.\u0026#160;\u0026#x21a9;\u0026#xfe0e;\nvan Eijk et al. (2022). Using the ALSFRS-R in multicentre clinical trials for amyotrophic lateral sclerosis: potential limitations in current standard operating procedures.\u0026#160;\u0026#x21a9;\u0026#xfe0e;\nvan Eijk et al. (2022). Joint modeling of endpoints can be used to answer various research questions in randomized clinical trials.\u0026#160;\u0026#x21a9;\u0026#xfe0e;\n","date":"11 May 2026","externalUrl":null,"permalink":"/blog/alsfrs-critical-appraisal-part-ii/","section":"Blog","summary":"A structured tour of the ALSFRS-R: its twelve items across four functional domains, why feasibility and sensitivity to change made it the field’s standard outcome measure for twenty-five years, and a frank account of the psychometric limitations that have accumulated against it.","title":"The ALSFRS-R: what it measures, why we use it, and where it fails","type":"blog"},{"content":"A talk presenting an update on the data collected in the Spanish ALS Registry and the analyses under way, given at the 3rd Meeting of the Spanish ALS Research Network (III Encuentro de la Red Española de Investigación en ELA; Zaragoza, 7–8 May 2026). The talk was part of the session \u0026ldquo;Current state of basic research: what we can do and what we cannot yet answer\u0026rdquo;, alongside talks on omics (Oriol Dols-Icardo) and disease models (Abraham Acevedo-Arozena).\nPresenting the update on the Spanish ALS Registry at the 3rd Meeting of the Spanish ALS Research Network (Zaragoza, 7 May 2026). The registry, run by the Fundación Francisco Luzón within the GENRARE framework, already gathers 3,136 patients entered by 125 active researchers across 52 hospitals in 14 autonomous communities. The talk covered:\nDescriptive characterisation and cohort-wide figures — growth of the number of patients included, distribution by site of onset and by phenotype, and analysis of cohort-wide figures for diagnostic delay and survival (the latter somewhat longer than the European PRECISION-ALS reference).1 Cognitive assessments — analysis of baseline ECAS scores, comparison across territories, and longitudinal analysis using linear mixed models. Effectiveness of riluzole — the registry data reproduce known findings: greater effectiveness in patients treated earlier, in those with a higher progression rate, and in those with respiratory onset. Real-world data on tofersen — stabilisation of the decline in ALSFRS-R scores when comparing the periods before and after treatment initiation in treated patients. Differences between autonomous communities — heterogeneity in progression rates across territories and in time to need for non-invasive ventilation, although data consistency must be established before attributing those differences to biological or care-related factors. It closed with the registry\u0026rsquo;s future directions: improving the statistical analysis, advancing the automated ingestion of clinical data, digging deeper into territorial differences, and interoperability with European platforms such as PRECISION ALS.\nThe meeting, organised by the Fundación Luzón at the Faculty of Veterinary Medicine of the University of Zaragoza, brought together close to a hundred researchers over two days around the causes and mechanisms of the disease, diagnosis and patient stratification, clinical trials and emerging treatments, as well as focus groups, poster sessions, and oral communications. It also marked the formal constitution of the Spanish ALS Research Network as an association of research groups, with its statutes and governance structure agreed collectively.\nGroup photo of the participants at the 3rd Meeting of the Spanish ALS Research Network (Faculty of Veterinary Medicine, University of Zaragoza). Caravaca Puchades et al. (2025). Mapping the natural history of amyotrophic lateral sclerosis: time-to-event analysis of clinical milestones in the pan-European, population-based PRECISION-ALS cohort.\u0026#160;\u0026#x21a9;\u0026#xfe0e;\n","date":"7 May 2026","externalUrl":null,"permalink":"/talks/3rd-meeting-spanish-als-research-network/","section":"Talks","summary":"An update on the data collected in the Spanish ALS Registry and the analyses under way, presented at the 3rd Meeting of the Spanish ALS Research Network (Zaragoza): descriptive characterisation and cohort-wide figures, cognitive assessments (ECAS), effectiveness of riluzole, real-world data on tofersen, and differences between autonomous communities.","title":"3rd Meeting of the Spanish ALS Research Network","type":"talks"},{"content":"","date":"7 May 2026","externalUrl":null,"permalink":"/authors/alejandro-caravaca-puchades/","section":"Authors","summary":"","title":"Alejandro Caravaca Puchades","type":"authors"},{"content":"","date":"7 May 2026","externalUrl":null,"permalink":"/authors/","section":"Authors","summary":"","title":"Authors","type":"authors"},{"content":"","date":"7 de May de 2026","externalUrl":null,"permalink":"/es/tags/ela/","section":"Tags","summary":"","title":"ELA","type":"tags"},{"content":"","date":"7 de May de 2026","externalUrl":null,"permalink":"/es/tags/evidencia-de-vida-real/","section":"Tags","summary":"","title":"Evidencia De Vida Real","type":"tags"},{"content":"","date":"7 May 2026","externalUrl":null,"permalink":"/tags/national-als-registry/","section":"Tags","summary":"","title":"National ALS Registry","type":"tags"},{"content":"","date":"7 May 2026","externalUrl":null,"permalink":"/tags/real-world-evidence/","section":"Tags","summary":"","title":"Real-World Evidence","type":"tags"},{"content":"","date":"7 de May de 2026","externalUrl":null,"permalink":"/es/tags/registro-nacional-ela/","section":"Tags","summary":"","title":"Registro Nacional ELA","type":"tags"},{"content":"A selection of conference presentations and invited talks.\n","date":"7 May 2026","externalUrl":null,"permalink":"/talks/","section":"Talks","summary":"","title":"Talks","type":"talks"},{"content":"Clinical trials in ALS live or die by their outcome measures. The choice of what to measure, how to measure it, and how to model the result determines whether a treatment effect is detectable, whether a regulatory submission is credible, and ultimately whether an effective drug reaches patients or languishes in a file cabinet \u0026mdash; is how we measure the consequences of that neuronal loss in living people over time.\nThis series is about that measurement problem. Specifically, it is about the Revised ALS Functional Rating Scale (ALSFRS-R): why it became the dominant outcome measure in ALS clinical trials, what it does well, what it does poorly, and what we might do instead \u0026mdash; or alongside it.\nThe measurement problem in neurodegeneration # A tumor shrinks or it doesn\u0026rsquo;t \u0026mdash; response is measured by imaging criteria that are unambiguous and widely accepted. The biology announces itself on the scan. In neurodegenerative disease, the situation is fundamentally different. The pathological process \u0026mdash; neuronal loss, protein aggregation, synaptic failure \u0026mdash; unfolds invisibly over years, while its functional consequences accumulate gradually and unevenly across biological systems. There is rarely a single biomarker that captures the whole picture. And unlike a shrinking tumour, functional decline in neurodegeneration has no single legible endpoint — it accumulates gradually, unevenly, and only becomes visible in aggregate.\nThis forces a choice. We can measure what is happening at the biological level: protein concentrations in cerebrospinal fluid, axonal loss on imaging, electrophysiological signatures of denervation. Or we can measure what is happening at the functional level: what the patient can and cannot do with their body on a given day. Both approaches have merit. Both have serious limitations. And the relationship between them \u0026mdash; between biological disease activity and functional disability \u0026mdash; is neither linear nor stable across individuals or time.\nA quantitative outcome measure is any instrument designed to translate that complexity into a number. The number is what goes into a statistical model. The statistical model is what informs a regulatory decision. Before committing to a given outcome measure, it is worth asking whether that number truly captures what we intend it to.\nWhat makes a good outcome measure? # Before evaluating any specific instrument, it helps to establish what we are looking for. A good outcome measure in a clinical trial should satisfy several criteria simultaneously, and they are in tension with each other.\nIt should be sensitive to change \u0026mdash; capable of detecting meaningful differences between treated and untreated patients within a feasible trial duration.\nIt should be reliable \u0026mdash; producing consistent results when administered by different raters, or in different settings, or at different times of day.\nIt should be valid \u0026mdash; actually measuring what it claims to measure, and not something adjacent to it.\nIt should exhibit interval-level measurement properties, meaning that a one-unit change at the bottom of the scale represents the same quantum of disease progression as a one-unit change at the top.\nAnd it should be feasible \u0026mdash; cheap enough, fast enough, and simple enough to be administered at scale across multi-site international trials.\nNo current outcome measure in ALS satisfies all of these criteria fully. The interesting question is which compromises are tolerable, and for whom.\nWhy not survival? # The most obvious endpoint in a fatal disease is survival. If a drug works, patients should live longer. If it doesn\u0026rsquo;t, they won\u0026rsquo;t. Survival is objective, clinically meaningful, and requires no rater training. It is also, in ALS, a deeply impractical primary endpoint for most trials.\nThe problem is time. Median survival in ALS from symptom onset is approximately 2–4 years, with enormous heterogeneity.1 Detecting a meaningful survival benefit \u0026mdash; say, a 20% reduction in mortality risk \u0026mdash; with adequate statistical power would require following thousands of patients for years. The resulting trials would be prohibitively expensive, logistically nightmarish, and ethically complicated. In an era where patients have access to compassionate use programmes, off-label treatments, and natural history registries, maintaining a clean placebo comparison over years of follow-up is increasingly untenable — patients and families will not accept it, and the regulatory environment does not require it. Survival also integrates everything \u0026mdash; disease biology, respiratory management, nutritional support, access to palliative care \u0026mdash; making it a noisy signal for the specific effect of a drug on neurodegeneration.\nFor these reasons, survival has largely been relegated to secondary endpoint status in ALS trials, used to support and contextualise results from faster-moving instruments rather than to anchor them.\nWhy not respiratory function, strength, or biomarkers? # The alternatives are plentiful, and each illuminates a different facet of the disease while obscuring others.\nForced vital capacity (FVC) tracks respiratory muscle decline with high precision and has strong prognostic relevance \u0026mdash; respiratory failure is the leading cause of death in ALS.2 But FVC is a single-domain measure. A drug that slowed limb function decline without affecting respiratory progression would be invisible to it, or nearly so. It also requires spirometry equipment and trained personnel, limiting its use in remote or resource-constrained settings.\nStrength testing \u0026mdash; whether by handheld dynamometry, megascore composite, or the older Medical Research Council grading \u0026mdash; offers direct measurement of motor output. The ATLIS system and related platforms have improved standardisation considerably.3 But strength tests are sensitive to effort, fatigue, and positioning, require in-person administration, and again measure only one dimension of a multisystem disease.\nBiomarkers are where the field\u0026rsquo;s hopes currently reside. Neurofilament light chain (NfL) in plasma and CSF reflects neuroaxonal damage with impressive sensitivity, correlates with disease progression, and changes in response to treatment in some contexts.4 But NfL is not yet a validated surrogate endpoint \u0026mdash; the regulatory bar for surrogate endpoints requires demonstrated correlation with clinical outcomes that withstands scrutiny across trials and populations, and that evidence base is still being built. Until that work is done, biomarkers function as supportive evidence rather than as the primary evidentiary standard.\nEnter the functional composite scale # Against this backdrop, the appeal of a functional composite scale becomes clear. If the disease affects bulbar function, upper limb function, lower limb function, and respiratory function \u0026mdash; and if no single domain captures the whole picture \u0026mdash; then a scale that samples across all of those domains simultaneously has an obvious advantage. It is sensitive to changes anywhere in the nervous system. It integrates information across biological subsystems. And if designed correctly, it can be administered quickly, reliably, and without expensive equipment.\nThis was the logic behind the original ALS Functional Rating Scale, published in 1996, and its revision in 1999 \u0026mdash; the ALSFRS-R \u0026mdash; which added greater granularity in the respiratory domain.5 Over the subsequent two and a half decades, it became the standard primary endpoint in ALS clinical trials worldwide: present in PRO-ACT, AnswerALS, and virtually every randomised controlled trial in the field. Whatever its flaws \u0026mdash; and they are substantial, which is the subject of the next three articles in this series \u0026mdash; it has the enormous practical advantage of being everywhere. Any new measure will be evaluated against it, and any analytical innovation that can be applied to existing ALSFRS-R data carries immediate leverage across an enormous accumulated evidence base.\nUnderstanding what the ALSFRS-R is, what it measures, and where it breaks down is therefore not merely an academic exercise. It is a prerequisite for designing better trials, interpreting existing ones, and \u0026mdash; ultimately \u0026mdash; getting effective treatments to patients faster.\nIn Part II, we examine the scale in detail: its structure, its strengths, and the long list of limitations that have accumulated in the literature over twenty-five years of use.\nLonginetti et al. (2019). Epidemiology of amyotrophic lateral sclerosis: an update of recent literature.\u0026#160;\u0026#x21a9;\u0026#xfe0e;\nBourke et al. (2006). Effects of non-invasive ventilation on survival and quality of life in patients with amyotrophic lateral sclerosis: a randomised controlled trial.\u0026#160;\u0026#x21a9;\u0026#xfe0e;\nAndres et al. (2012). Validation of a new strength measurement device for amyotrophic lateral sclerosis clinical trials.\u0026#160;\u0026#x21a9;\u0026#xfe0e;\nBenatar et al. (2023). Neurofilament light chain in drug development for amyotrophic lateral sclerosis: a critical appraisal.\u0026#160;\u0026#x21a9;\u0026#xfe0e;\nCedarbaum et al. (1999). The ALSFRS-R: a revised ALS functional rating scale that incorporates assessments of respiratory function.\u0026#160;\u0026#x21a9;\u0026#xfe0e;\n","date":"5 May 2026","externalUrl":null,"permalink":"/blog/alsfrs-critical-appraisal-part-i/","section":"Blog","summary":"Clinical trials in ALS live or die by their outcome measures. This opening piece of the series frames the measurement problem in neurodegeneration — biological versus functional endpoints — and sets up why the ALSFRS-R became the field’s default answer to it.","title":"What are we actually measuring?","type":"blog"},{"content":"","date":"4 de June de 2025","externalUrl":null,"permalink":"/es/tags/cognici%C3%B3n/","section":"Tags","summary":"","title":"Cognición","type":"tags"},{"content":"","date":"4 June 2025","externalUrl":null,"permalink":"/tags/cognition/","section":"Tags","summary":"","title":"Cognition","type":"tags"},{"content":"A platform presentation given at the ENCALS Meeting 2025 (Turin, Italy, 3–6 June 2025; Centro Congressi Lingotto), in a plenary session on clinical characterisation, genetics and prognostic factors in ALS. The talk — \u0026ldquo;Examining the Cognitive Profile of Restricted Phenotypes in ALS: Insights from the National Registry of Ireland\u0026rdquo; — presented work carried out with the Academic Unit of Neurology at Trinity College Dublin (Robert McFarlane, Emmet Costello, Éanna Mac Domhnaill, Mark Heverin and Orla Hardiman), drawing on the National ALS Register of Ireland, during a research rotation at Trinity College Dublin.\nPresenting the cognitive profile of restricted ALS phenotypes at the ENCALS Meeting 2025 (Turin, 4 June 2025). Most people with ALS decline over a couple of years; a few do not — and the restricted phenotypes are the clearest example. Progressive bulbar palsy (PBP), flail arm syndrome (FAS) and flail leg syndrome (FLS) are forms of the disease that stay confined to one region for far longer and carry a markedly better prognosis. In other words, these patients already have what every ALS therapy is trying to deliver — a brake on progression. That is what makes them worth studying in their own right: whatever underlies their slower course is a clue to how the disease might be slowed in everyone else. The talk looked at who these patients are and whether their motor stability is matched by their cognition, using the National ALS Register of Ireland. In outline:\nA data-driven definition of \u0026ldquo;restricted\u0026rdquo; disease. King\u0026rsquo;s stage was estimated from the data available in the register, and phenotype-specific cut-offs for how long a patient stays in the earliest stage were set by a landmark-based threshold analysis of survival — so \u0026ldquo;restricted\u0026rdquo; is read off the clinical record rather than assigned by label alone, with the threshold tailored to each of PBP, FAS and FLS. What sets restricted patients apart. Compared with the rest of the cohort, restricted patients tended to be younger at onset, to progress more slowly on functional measures, and to have had a longer diagnostic delay — with the sex balance and the length of that delay differing across the three subtypes. Cognition. On cross-sectional ECAS scores, restricted patients looked cognitively better off than the rest of the cohort, with fewer of them falling in the impaired range; and when longitudinal ECAS scores were modelled with generalised additive mixed models (GAMMs), they not only started from a higher level but showed evidence of greater stability over time than the unrestricted phenotypes. The picture, then, is of patients whose disease is held back on both the motor and the cognitive side — and the open question, the one that motivates the work, is what does the holding back.\nDuring the plenary session at the ENCALS Meeting 2025 (Turin, June 2025). The ENCALS Meeting is the annual conference of the European Network to Cure ALS; the 2025 edition brought roughly a thousand researchers and clinicians to Turin over four days for satellite meetings, plenaries, platform sessions and posters spanning ALS genetics, biomarkers, clinical trials and care.\nThe auditorium at the Centro Congressi Lingotto during the ENCALS Meeting 2025 (Turin). ","date":"4 June 2025","externalUrl":null,"permalink":"/talks/encals-meeting-2025/","section":"Talks","summary":"A plenary-session oral communication at the ENCALS Meeting 2025 (Turin) on the cognitive profile of restricted ALS phenotypes — progressive bulbar palsy, flail arm and flail leg syndromes — using the National ALS Register of Ireland: why these slow-progressing forms are worth studying, a data-driven way to define “restricted” disease from the clinical record, what distinguishes these patients, and ECAS analyses pointing to comparatively stable cognition over time.","title":"ENCALS Meeting 2025","type":"talks"},{"content":"","date":"4 de June de 2025","externalUrl":null,"permalink":"/es/tags/fenotipos/","section":"Tags","summary":"","title":"Fenotipos","type":"tags"},{"content":"","date":"4 June 2025","externalUrl":null,"permalink":"/tags/irish-als-register/","section":"Tags","summary":"","title":"Irish ALS Register","type":"tags"},{"content":"","date":"4 June 2025","externalUrl":null,"permalink":"/tags/phenotypes/","section":"Tags","summary":"","title":"Phenotypes","type":"tags"},{"content":"","date":"4 de June de 2025","externalUrl":null,"permalink":"/es/tags/registro-de-ela-de-irlanda/","section":"Tags","summary":"","title":"Registro De ELA De Irlanda","type":"tags"},{"content":" Clinical work # I am a neurologist working in clinical neurology with a focus on amyotrophic lateral sclerosis (ALS) and other motor neuron diseases. My day-to-day practice involves the diagnosis, follow-up, and multidisciplinary care of patients living with ALS, alongside collaboration with allied specialties — pulmonology, nutrition, rehabilitation, and palliative care — that shape what good ALS care actually looks like.\nData science and quantitative work # Alongside clinical practice, I work as a data scientist on questions that arise from the same patients I see in clinic. My quantitative work centres on registry-based research, longitudinal modelling of disease progression, survival analysis, and the methodological problems that come with messy real-world data: missingness, censoring, measurement variability, and the gap between trial populations and the people who actually walk into the clinic. I am especially interested in the intersection between epidemiology, biostatistics, and software — the parts of the pipeline where bad tooling silently produces bad evidence.\nCurrent projects # PRECISION ALS — A pan-European research programme building a federated ALS patient data platform across nine research centres, in partnership with the TRICALS consortium, to support large-scale real-world evidence generation toward new treatments. National ALS Registry of Spain — I serve as national data manager, coordinating data structure, quality control, and analytical workflows across participating Spanish centres. ","externalUrl":null,"permalink":"/about/","section":"About","summary":"","title":"About","type":"about"},{"content":"","externalUrl":null,"permalink":"/categories/","section":"Categories","summary":"","title":"Categories","type":"categories"},{"content":"A current version of my CV is available as a PDF.\n","externalUrl":null,"permalink":"/cv/","section":"CV","summary":"","title":"CV","type":"cv"},{"content":"A list of peer-reviewed articles and conference proceedings, grouped by year. Click on a title to follow its DOI; expand the abstract to read more.\n","externalUrl":null,"permalink":"/publications/","section":"Publications","summary":"","title":"Publications","type":"publications"},{"content":"A list of open-source software projects I maintain. Click on a title to visit the source repository.\n","externalUrl":null,"permalink":"/software/","section":"Software","summary":"","title":"Software","type":"software"}]