At‐Home Versus in‐Clinic Vital Capacity Measurement: Insights From the HEALEY ALS Platform Trial Authors: Eric A Macklin, James D Berry, Brittney A Harkey, Lindsay Heyd, Marianne Chase, Hong Yu, Alexander V Sherman, Derek Dagostino, Gale Kittle, Meghan Hall, Mariah R Connolly, Kristin Drake, Elisa Giacomelli, Erica Scirocco, Saloni Sharma, Suma Babu, Michael Benatar, Zachary Simmons, Eufrosina Young, Merit E Cudkowicz, Jeremy Shefner, Sabrina Paganoni, the HEALEY ALS Platform Trial Study Group Published: 2026 ABSTRACT ABSTRACTIntroduction/AimsRespiratory weakness, typically monitored as vital capacity (VC), is a central feature of amyotrophic lateral sclerosis (ALS). VC is increasingly measured remotely in participants' homes, although in‐clinic assessment remains the standard. We tested concordance between at‐home and in‐clinic VC to determine trial eligibility, track progression, and predict survival in a large ALS trial.MethodsAt‐home and in‐clinic VC were assessed at baseline and approximately every 8 weeks for a year in the first four regimens of the HEALEY ALS Platform Trial. At‐home assessments were coached via live videoconference and centrally reviewed. VC measurements, expressed as percent of predicted normal (%PN), were compared cross‐sectionally, longitudinally, and for predicting survival time. Data from 233 participants with 3–8 paired at‐home and in‐clinic VC assessments completed < 14 days apart were analyzed.ResultsAt‐home and in‐clinic VC were well correlated (Lin's r c = 0.82) with no systematic bias. At‐home VC ≥ 60%PN predicted in‐clinic VC ≥ 60%PN with a positive predictive value of 91% and a negative predictive value of 65%. VC slopes were moderately correlated (r c = 0.68). At‐home VC progressed 28% faster than in‐clinic VC with proportionately less variance (at‐home [SE] = −1.882 [0.153] %PN/month, in‐clinic = −1.476 [0.131] %PN/month). Slopes of at‐home and in‐clinic VC explained 15% and 17% of variation in future survival time, respectively.DiscussionAt‐home and in‐clinic VC were well correlated cross‐sectionally. At‐home VC performed well tracking longitudinal change and predicting survival. The reduced participant burden of assessment and concordance with in‐clinic measurement support use of at‐home monitoring of VC. Keywords: amyotrophic lateral sclerosis, remote patient monitoring, spirometry, validation studies as topic Introduction Respiratory weakness is a hallmark of amyotrophic lateral sclerosis (ALS). Vital capacity (VC) is the primary measure of respiratory function used for monitoring ALS progression [1]. Baseline VC and change in VC predict survival in ALS [2, 3, 4]. In‐clinic respiratory function testing is the gold standard for measuring VC. Remote measurement of VC has been used since the 1990's and gained importance during the COVID‐19 pandemic. Prior work has reported good correlation between at‐home and in‐clinic VC among people living with ALS [5, 6]. Data on agreement between at‐home and in‐clinic assessments for monitoring change in VC or predicting survival are lacking. This study uses data from the HEALEY ALS Platform Trial to evaluate the accuracy and utility of at‐home VC as a proxy for in‐clinic measurements. Methods The HEALEY ALS Platform Trial [7, 8, 9] (NCT04297683) is a perpetual multi‐center, multi‐regimen clinical trial evaluating the safety and efficacy of investigational products for the treatment of ALS. Eligibility criteria included adults with revised El Escorial criteria of possible, laboratory‐supported probable, probable, or definite sporadic or familial ALS; onset of weakness < 36 months before screening; slow vital capacity (SVC) ≥ 50%‐predicted by GLI‐2012 [10]; no use or on a stable dose of riluzole, edaravone, or both; no unstable clinical or psychiatric conditions in investigator's opinion; and no recent exposure to other investigational treatments or prior gene therapy. Participants completed in‐clinic visits at baseline and 4, 8, 16, 24, 28, 32, 40, and 48 or 52 weeks. The master protocol and regimen‐specific appendices were approved by FDA and IRB, and participants provided written informed consent. This study utilizes data from the first four regimens [11, 12, 13, 14]. SVC was assessed at in‐clinic visits (excluding week 4 and visits during the pandemic when aerosol‐generating procedures were restricted) using the EasyOne Air spirometer (ndd Medical Technologies, Andover, MA). Forced vital capacity (FVC) was assessed remotely using the MIR Spirobank Smart spirometer (Medical International Research, Rome, Italy) with data collection and video interface provided by the ZEPHYRx smartphone app (ZEPHYRX LLC, Troy, NY). In each case, certified evaluators coached participants through 3 to 5 maneuvers. A minimum of 2 completed maneuvers was required for analysis (69 of 3996 assessments [1.7%] were excluded). For at‐home assessments, central readers at the Barrow Neurologic Institute Outcomes Center reviewed flow‐loops and adjudicated which assessments were acceptable (235 of 2460 coached assessments [9.6%] were excluded). ATS/ERS 2019 standards [15] were not referenced. The maximum expired volume from accepted assessments was converted to percent of predicted normal (%PN) using GLI‐2012 norms [10]. Pairs of acceptable at‐home and in‐clinic VC measurements were included for analysis if they were assessed fewer than 14 days apart. Participants with paired assessments for at least 3 of 8 planned study visits were included to support comparison of longitudinal changes in VC. Date of death or date last known alive was recorded by sites at the conclusion of each participant's follow‐up as well as for all participants at the conclusion of the last participant's follow‐up and by a forensic service using publicly available data biannually after the last participant's follow‐up. All paired at‐home and in‐clinic VC assessments were compared using Bland–Altman plots. Within‐person, cross‐sectional correlation between at‐home and in‐clinic VC was estimated as Lin's concordance correlation coefficient (rc) from a bivariate linear mixed model with unstructured participant‐specific random effects over assessment methods and unstructured participant‐specific repeated measures over visits [16]. Operating characteristics for using at‐home VC ≥ 60%PN as a proxy for in‐clinic VC ≥ 60%PN as a typical inclusion criterion were estimated using mixed model logistic regression with random participant‐specific intercepts on the log‐odds scale. Rates of change of at‐home and in‐clinic VC were estimated in separate linear mixed models with a fixed effect of time and unstructured participant‐specific random intercepts and slopes. Relative precision of slopes was calculated as the estimated slope divided by the slope standard error and multiplied by the square root of the sample size. Prognosis of survival beyond each participant's final follow‐up visit was estimated in separate Cox proportional hazards regressions with and without adjustment for baseline neurofilament light chain protein (NfL) concentration (log‐transformed) and TRICALS score [17]. Explained variation [18] was calculated for each model and as an increment versus a model with NfL and TRICALS only. Analyses were performed using SAS v9.4 (SAS Institute, Cary, NC). Results Data were included from 233 participants enrolled from August 2020 to July 2022 out of 439 who completed at least one acceptable at‐home assessment. Others were excluded due to ≥ 14‐day gaps in assessment or < 3 paired assessments. Mean follow‐up was 316 days (SD 71, range 154 to 406 days). Vital status data were collected through January 2026. Participants were generally early in disease course (Table 1). Due to exclusion of participants who died or were lost to follow‐up prior to completing at least 3 paired assessments, the analysis cohort was enriched for slower progressors (Table S1). TABLE 1Baseline characteristics.CharacteristicIn‐clinic SVC and At‐home FVC < 14 days apart at 3 to 8 visits (n = 233)SexMale149 (63.9%)RaceAsian4 (1.7%)Black or African American3 (1.3%)White222 (95.3%)Unknown/not reported4 (1.7%)EthnicityHispanic or Latino7 (3.0%)Age at baseline (yrs)Mean ± SD (Range)59.4 ± 10.6 (23.6, 81.9)Body mass index (kg/m2)Mean ± SD (Range)26.7 ± 4.8 (16.3, 47.0)Site of symptom onsetBulbar30 (12.9%)Revised El Escorial criteriaPossible17 (7.3%)Probable Lab‐supported61 (26.2%)Probable91 (39.1%)Definite64 (27.5%)Riluzole use at baselineYes189 (81.1%)Edaravone use at baselineYes53 (22.7%)Nuedexta use at baselineYes25 (10.7%)Months since symptom onsetMean ± SD (Range)22.0 ± 8.5 (2.7, 37.7)Months since diagnosisMean ± SD (Range)11.4 ± 7.2 (1.8, 33.1)Delta‐FRS (pts/month)Mean ± SD (Range)0.62 ± 0.39 (0.00, 2.09)TRICALS linear predictor scoreMean ± SD (Range)−4.46 ± 1.28 (−9.01, −1.39)Serum NfL at baseline (ng/L)Median (Q1‐Q3)72.0 (46.8104)ALSFRS‐R Total scoreMean ± SD (Range)36.0 ± 6.0 (14.0, 48.0)ALSFRS‐R Respiratory domainMean ± SD (Range)10.5 ± 2.3 (1.0, 12.0)In‐clinic SVC (%‐predicted)Mean ± SD (Range)80.7 ± 17.2 (33.9163)At‐home FVC (%‐predicted)Mean ± SD (Range)82.1 ± 18.5 (23.3128) Baseline characteristics. At‐home and in‐clinic VC were well correlated (Figure 1A) with no systematic bias but wide limits of agreement (Figure 1B). Good correlation was observed regardless of bulbar function (Table S2). At‐home VC measurements ≥ 60%PN had good overall accuracy and positive predictive value but only moderate negative predictive value (Table S3). Scatterplots (A, C) and Bland–Altman plots (B, D) comparing in‐clinic SVC and at‐home FVC estimates cross‐sectionally (A, B) and longitudinally (C, D). All units are %‐predicted. All assessments were completed < 14 days apart and are reported as percentage of predicted normal values. r[c] = Lin's concordance correlation coefficient, LOA, limit of agreement. Slopes of at‐home and in‐clinic VC were moderately correlated (Figure 1C). At‐home VC progressed 28% faster across the range of VC estimates (Figure 1D) with proportionately less variance, yielding a standardized slope 9% larger than in‐clinic VC (Table S4). Assuming equal proportional effect of an intervention on change in VC by either assessment method, equivalent follow‐up schedules and missed visits, and a comparable patient population to that studied in this analysis, trials using at‐home instead of in‐clinic VC assessments would require sample sizes 17% smaller. Slopes of at‐home VC estimated over 48–52 weeks explained 90% of the variation explained by slopes of in‐clinic VC (Table 2). Both VC methods explained variation in future survival time beyond that explained by baseline NfL and TRICALS score to a similar degree. TABLE 2Prediction of future survival.EndpointCovariatePrognosticHazard ratio a (95% CI)Chi‐squared statisticExplained variation b Additional explained variation c In‐clinic SVC (%‐predicted)UnadjustedLast value0.870 (0.797, 0.949)9.716.65%NASlope0.861 (0.812, 0.914)24.3Baseline NfL and TRICALSLast value0.922 (0.841, 1.009)3.123.38%9.51%Slope0.875 (0.822, 0.931)17.6At‐home FVC (%‐predicted)UnadjustedLast value0.915 (0.847, 0.989)5.114.90%NASlope0.844 (0.786, 0.907)21.7 Baseline NfL and TRICALS Last value0.955 (0.883, 1.035)1.322.01%8.14%Slope0.860 (0.797, 0.929)14.8 a The hazard ratios for the person‐specific last value of each endpoint reflect the magnitude of reduction in hazard associated with a difference of 10%‐predicted. The hazard ratios for the person‐specific slope of each endpoint estimated from a random‐slopes linear mixed model reflect the magnitude of reduction in hazard associated with a slowing of progression equal to 30% of the mean progression rate for each endpoint. b Variation in event times explained by the full model, including in‐clinic SVC or at‐home FVC measures alone for unadjusted models or also with baseline log‐transformed serum NfL concentration and TRICALS score as covariates. c Variation in event times explained by in‐clinic SVC or at‐home FVC measures beyond that explained by baseline log‐transformed serum NfL concentration and TRICALS score. Prediction of future survival. Baseline NfL and TRICALS The hazard ratios for the person‐specific last value of each endpoint reflect the magnitude of reduction in hazard associated with a difference of 10%‐predicted. The hazard ratios for the person‐specific slope of each endpoint estimated from a random‐slopes linear mixed model reflect the magnitude of reduction in hazard associated with a slowing of progression equal to 30% of the mean progression rate for each endpoint. Variation in event times explained by the full model, including in‐clinic SVC or at‐home FVC measures alone for unadjusted models or also with baseline log‐transformed serum NfL concentration and TRICALS score as covariates. Variation in event times explained by in‐clinic SVC or at‐home FVC measures beyond that explained by baseline log‐transformed serum NfL concentration and TRICALS score. Discussion At least four features distinguished the at‐home and in‐clinic VC assessment methods used in this study: (1) the setting and ability of the evaluator to facilitate a good seal on the spirometer and to elicit maximal effort, (2) the use of forced versus slow maneuvers, (3) the use of different spirometers, and (4) the use or non‐use of central review for qualifying flow‐loops. Nevertheless, we observed strong cross‐sectional correlation between the two methods with almost no bias on average and limited difference in scale, consistent with previously published studies [5, 6]. Cross‐sectional correlation was strong. Nevertheless, there were wide limits of agreement, and at‐home VC underestimated in‐clinic VC at the low end of the scale. In our population, this underestimation could exclude roughly 35% of otherwise eligible participants from trials based on in‐clinic VC. However, NPV is dependent upon population characteristics and would improve in a population with more individuals with VC substantially below the required cutoff. At‐home VC progressed 28% faster on average than in‐clinic VC over 48 to 52 weeks in our cohort, but slopes of the two measures were moderately correlated. Although progression rates of at‐home VC were faster, comparable explained variation in long‐term survival suggests that at‐home and in‐clinic VC trends reflected underlying disease progression with similar accuracy. Moreover, comparable precision estimating at‐home and in‐clinic VC slopes suggests that at‐home VC could be an equally powerful endpoint when tracked over a year. This study had limitations. Our comparison of at‐home and in‐clinic VC was confounded by use of different maneuvers and devices. Prior work has shown good correlation between FVC and SVC in ALS patients cross‐sectionally during in‐clinic assessment [19, 20] and longitudinally during remote assessment [21], although systematically higher values were reported for SVC among patients with bulbar dysfunction [22]. We are not aware of studies comparing the EasyOne and MIR spirometers. Results reported here may not generalize to other ALS populations. The exclusion of participants with fewer than 3 paired at‐home and in‐clinic assessments yielded a sample with slower disease progression, better baseline functional status, and higher vital capacity than the overall HEALEY ALS Platform Trial. There may also be an unmeasured bias toward people who were able to successfully use the at‐home spirometer. Agreement between at‐home and in‐clinic VC would likely decline over shorter intervals where the effect of residual variation is proportionally greater. The lower burden of at‐home VC assessments for participants and potential for participation of a more representative sample of people living with ALS are attractions. Quantifying the relative performance of at‐home versus in‐clinic VC permits informed decisions about the use of at‐home versus in‐clinic VC for measuring respiratory function at a point in time or for tracking changes in respiratory function to monitor disease progression. Our data, consistent with previous studies, suggest that at‐home VC, when performed with live video coaching and secondary central review and adjudication of all flow‐loops, provides a reasonable approximation of in‐clinic VC, both cross‐sectionally and longitudinally. The reduced burden of assessment for participants, the potential for more frequent assessment, the cross‐sectional and prognostic accuracy, and the longitudinal correlation and precision support the use of at‐home VC measurement both clinically and in ALS trials. Author Contributions Brittney A. Harkey: resources. Lindsay Heyd: resources. Gale Kittle: resources, methodology. Alexander V. Sherman: resources. Eric A. Macklin: conceptualization, investigation, writing – original draft, methodology, writing – review and editing, visualization, validation. Kristin Drake: resources. Hong Yu: resources. Saloni Sharma: writing – review and editing. Marianne Chase: resources. Derek Dagostino: resources. James D. Berry: conceptualization, investigation, writing – original draft, methodology, writing – review and editing. Zachary Simmons: resources, writing – review and editing. Suma Babu: writing – review and editing. Jeremy Shefner: supervision, conceptualization, writing – review and editing, methodology. Erica Scirocco: writing – review and editing. Sabrina Paganoni: conceptualization, supervision, writing – review and editing. Elisa Giacomelli: writing – review and editing. Mariah R. Connolly: resources. Michael Benatar: resources, writing – review and editing. Merit E. Cudkowicz: writing – review and editing, resources. Eufrosina Young: resources, writing – review and editing. Meghan Hall: resources. Funding This work was supported by the Muscular Dystrophy Association (MDA 1313867), Tackle ALS, ALS Finding a Cure, ALS Association, ALS ONE, The Arthur M. Blank Family Foundation, the AMG Charitable foundation. Disclosure Suma Babu has received research funding from OrphAI therapeutics, Biogen, Denali Therapeutics, Novartis, Orion, Ionis, American Academy of Neurology, AANEM Foundation, and Muscular Dystrophy Association. She has received other honoraria from Biogen, Medscape, and MDA. Michael Benatar reports consulting fees from Alaunos, Alector, Alexion, Amgen, Annexon, Arrowhead, Biogen, Bristol Myers Squibb, Canopy, Cartesian, CorEvitas, Denali, Eli Lilly, Immunovant, Janssen, Merck, Novartis, Prilenia, Roche, Sanofi, Takeda, UCB, uniQure, Voyager, and Woolsey. The University of Miami has licensed intellectual property to Biogen to support the design of the ATLAS study. Michael Benatar is an unpaid member of the Board of Trustees of the ALS Association. James Berry reports consulting fees from Biogen, MT Pharma of America, MT Pharma Holdings of America, Janssen, ALexion, Regeneron, and Amylyx. He has received research support from Biogen, Clene Nanomedicine, MT Pharma of America, MT Pharma Holdings of America, Alexion, Amylyx, Rapa Therapeutics, Brainstorm Cell Therapeutics, MDA, ALSA, ALS Finding A Cure, ALS One, Tambourine, DoD, and NINDS. He has been a paid educational speaker for PeerView and Projects in Knowledge. Merit Cudkowicz reports research grants from Prilenia, Seelos, Calico, Denali, ITB, NIH, ALS Finding a Cure, and ALS One, and reports consulting fees from Pasithea, Transposon, Vector Y, Infectis, Roche, AB Science, Novartis, Regeneron, Quralis, ImmunityPharm, and Coya. She is a board member of Praxis Precision Medicine. Eric Macklin reports research support via his institution from Calico, Denali, HillHurst, ITB‐Med, Janssen, Eli Lilly, Mitsubishi Tanabe Pharma America, Neurizon, Prilenia, Transposon, Woolsey, Farmer Family Foundation, Michael J Fox Foundation, Muscular Dystrophy Association, and NIH and personal compensation from serving on Data Monitoring Committees for Argenx and Neurosense, from serving on Steering Committees for UCB, and from serving on advisory boards for Genentech and Merck. Sabrina Paganoni reports research grants from Amylyx Therapeutics, Revalesio, Eledon, Alector, UCB Pharma, Biohaven, Clene Nanomedicine, Prilenia Therapeutics, Seelos, Calico, Denali, Neurizon, NIH, DoD, the Muscular Dystrophy Association and reports consulting fees from Amylyx, Arrowhead, Biogen, BMS, Clene, Cytokinetics, Eikonizo, J&J, Merck, PharmAust, Prilenia, and Sola. She has been a paid educational speaker for Medscape, PeerView, and i3Health. Jeremy Shefner reports research funding from: NIH, AB Sciences; Acorda Therapeutics; Alector; Amylyx; Biogen; Cytokinetics Incorporated; Ionis; Mitsubishi Tanabe Pharma America; Quralis; PTC; Sanofi; Wave; Myolex, and personal compensation for consulting from: Amylyx [ALS]; Cytokinetics [ALS]; Denali [ALS]; GSK [ALS]; Mitsubishi Tanabe Pharma America [ALS]; Neurosense [ALS]; Orthogonal [ALS]; Pinteon [ALS]; RRD [ALS]; Acurastem [ALS]; Revalasio [ALS]; Apellis [ALS]; Swanbio [ALD]; Novartis [ALS]; Sanofi [ALS]; Eikonizo [ALS]; Sarepta [Muscular Dystrophy]. Alex Sherman reports research grants from The ALS Association, ALS Finding a Cure, and the NIH. Zachary Simmons, one of the co‐authors of this manuscript, currently serves as the Editor‐in‐Chief of Muscle & Nerve. To avoid any conflicts of interest, editorial handling and peer review of this manuscript were managed independently by another member of the editorial board. Eufrosina Young reports research support via her institution from The ALS Association and Mitsubishi Tanabe Pharma America. Ethics Statement We confirm that we have read the Journal's position on issues involved in ethical publication and affirm that this report is consistent with those guidelines. Conflicts of Interest The authors declare no conflicts of interest. Supporting information Table S1: Baseline characteristics comparing participants included in versus excluded from the primary analysis set. Table S2: Correlation between at‐home FVC and in‐clinic SVC assessed < 14 days apart for the full study cohort and stratified by concurrent ALSFRS‐R bulbar domain score, grouped to achieve similar sample sizes. Table S3: Prevalence of in‐clinic SVC ≥ 60%‐predicted and operating characteristics using at‐home FVC ≥ 60%‐predicted to predict eligibility based on in‐clinic SVC ≥ 60%‐predicted. Table S4: Estimated progression rates over 48–52 weeks. Figure S1: (A) Scatterplot of at‐home FVC estimates versus in‐clinic SVC. Dashed lines indicate threshold for hypothetical eligibility criterion ≥ 60%‐predicted. (B) Receiver‐operator characteristic (ROC) curve for prediction of in‐clinic VC ≥ 60%‐predicted by concurrent at‐home VC. C. Positive predictive value (PPV) versus negative predictive value (NPV) for prediction of in‐clinic VC ≥ 60%‐predicted by concurrent at‐home VC. (D) Precision‐recall curve for prediction of in‐clinic VC ≥ 60%‐predicted by concurrent at‐home VC. Note that the y‐axis of panels C and D begin at 0.80. Acknowledgments This investigation was supported by philanthropic gifts to the Sean M. Healey and AMG Center for ALS, by a grant from the Muscular Dystrophy Association (MDA award number: MDA 1313867), and by Tackle ALS, ALS Finding a Cure, The ALS Association, ALS ONE, The Arthur M. Blank Family Foundation, and The AMG Charitable Foundation. UCB, Biohaven Pharmaceuticals Inc., Clene Nanomedicine, and Prilenia Therapeutics provided investigational product and partially supported their individual regimens. We acknowledge the altruism of the participants and their families and the contributions of the HEALEY ALS Platform Trial research and support staff at each of the participating sites. We are grateful for the formative input we received from all members of the HEALEY ALS Platform Trial committee members past and present. Additionally, we sincerely appreciate the members of the Data and Safety Monitoring Board (Jason Connor, PhD; Borje Darpo, MD, PhD FESC; Karl Kieburtz, MD, MPH; Ruth Schneider, MD; and Andrew Siderowf, MD) for their oversight and guidance in ensuring the integrity of the trial and the safety of the participants. Macklin E. A., Berry J. D., Harkey B. A., et al., “At‐Home Versus in‐Clinic Vital Capacity Measurement: Insights From the HEALEY ALS Platform Trial,” Muscle & Nerve 74, no. 3 (2026): 702–708, 10.1002/mus.70322. Jeremy Shefner and Sabrina Paganoni co‐last authors. Contributor Information Eric A. Macklin, Email: emacklin@mgh.harvard.edu. the HEALEY ALS Platform Trial Study Group: Eric A. Macklin, Lori B. Chibnik, Douglas Hayden, Po‐Ying Lai, Rachel A. Donahue, Hao‐Wun Chen, Jianing Wang, Melanie Quintana, Benjamin R. Saville, Michelle A. Detry, Joe Marion, Matteo Vestrucci, Anna McGlothlin, Gustavo Alameda, Nithya Mathai, Eduardo Locatelli, Gabriela Lopes, Sabrina Paganoni, Merit Cudkowicz, Doreen Ho, Alexandra McCaffrey, James D. Berry, Suma Babu, Jennifer Scalia, Sarah Luppino, Clotilde Lagier‐Tourenne, Ghazaleh Sadri‐Vakili, Adam Quick, Stephen Kolb, Sarah Heintzman, Senda Ajroud‐Driss, Robert Sufit, April Szymanski, Jonathan Katz, Liberty Jenkins, Daragh Heitzman, Alan Martin, Stanley H. Appel, Simpson) Greene, Jason R. Thonhoff, Sheetal Shroff, Bing Liao, Kevin Felice, Charles Whitaker, Nicholas J. Maragakis, Lora L. Clawson, Alpa Uchil, Kristen M. Riley, JinAe Arneklev, Zachary Simmons, James Grogan, Xiaowei Su, Mansoureh Mamarabadi, Timothy M. Miller, Amber Malcolm, Nicholas Olney, Tracy Bazan, Michael D Weiss, Nassim Rad, Leo H. Wang, Stephen A. Goutman, Eva L. Feldman, Joseph Americo Fernandes, Ezequiel Piccione, Pariwat Thaisetthawatkul, Omar Jawdat, Constantine Farmakidis, Duaa Jabari, Jeffrey Statland, Mamatha Pasnoor, Mazen Dimachkie, Margaret Ayo Owegi, Robert H. Brown, Mehdi Ghasemi, Hajar Houmani, Catherine Douthwright, Kate Daniello, Laura A. Foster, Tuan Vu, Niraja Suresh, Jerrica Farias, I‐Hweii A. Chen, Hristelina Ilieva, Piera Pasinelli, Daniel S. Newman, Ximena Arcila‐Londono, Kara Steijlen, Carlayne E. Jackson, Ratna Bhavaraju‐Sanka, Shafeeq Ladha, Bill Jacobsen, Jourdan Milliard, Robert Bowser, Jeremy M. Shefner, Terry Heiman‐Patterson, Anahita Deboo, James B. Caress, Michael S. Cartwright, Andrea Swenson, Christopher Nance, Ludwig Gutmann, Jinsy Andrews, Julia Yasek, Matthew Harms, Amanda Peltier, Richard Lewis, Matthew Burford, Frank Diaz, Dominic Fee, David Shrilla, Matthew Elliott, Goran Rakocevic, Sarah Jones, Guillermo Solorzano, Richard Bedlack, Xiaoyan Li, Edward J. Kasarskis, Zabeen Mahuwala, Vishakhadatta (Vish) Mathur Kumaraswamy, Lauren Elman, Colin Quinn, Michael Baer, Jeffrey Rosenfeld, David Borg, Karthikeyan Bhuvaneswaran, Jasdeep Kaur, David Walk, Sam Maiser, Seward B. Rutkove, Courtney E. McIlduff, Paul Twydell, Andrew Mundwiler, Eufrosina Young, Jenny A. Meyer, Kristin Johnson, Pooja Rao, Kourosh Rezania, Betty Soliven, Raymond Roos, Namita A. Goyal, Ali A. Habib, Tahseen Mozaffar, Manisha Kak Korb, Jeffrey Mullen, Jeffrey A. Cohen, Elijah Stommel, Nathaniel M Robbins, Michael Benatar, Nathan Carberry, Volkan Granit, Raghav Govindarajan, Vovanti Jones, Jonathan Glass, Christina N. Fournier, Jaimin S. Shah, Bjorn Oskarsson, Said R. Beydoun, Leila Darki, Rodrigo Rodriguez, James P. Wymer, Miguel Chuquilin, Whitney McNeely, Lindsay Zilliox, Montserrat Diaz‐Abad, Peter H. Jin, Chandana Chauhan, Shakti Nayar, Gary L. Pattee, James Bobenhouse, Jennifer Martinez‐Thompson, Nathan P. Staff, Ghazala Hayat, Luisa Arroyave, Abbey Bailey, Jesse Bailey, Victoria Barlow, Allison Bulat, Genevive Changkuon, Marianne Chase, Melissa Cirino, Derek Dagostino, Cristina Deignan, Emma Deirmendjian, Annette De Mattos, Sofia DiStefano, Kristin Drake, Michaela Estes, Kenneth Faulconer, Precious Figueroa‐Szostek, Tessa Garozzo, Brittney A. Harkey, Meredith Gibbons Hasenoehrl, Jennifer Henrique, Natalie Henrique, Samuel Hurwitz, Courtney Igne, Liam Irwin, Katie Jentoft, Boglarka Jordan, Igor Katsovskiy, Olga Kharakozova, Taylor Kolvek, Alexander Korin, Thuong La, Haining Li, Joey Nguyen, Ilya Novak, Ricardo Ortiz, Joe Ostrow, Jaclyn Pagliaro, Jack Palillo, Payal Patel, Janae Patterson, Minh Phan, Najla Popel, Lindsay Pothier, Serena Proueng, Jesse Rosenthal, Alexander V. Sherman, Catherine Small, Natalia Tarasenko, Mirna Thomas, Eric Tustison, Prasha Vigneswaran, Yusra Wahab, Isaac Whitworth, Spencer Wright, Hong Yu, Mariah Connolly, Diana De Santiago, Adrian Felix, Karly Garrett, Meghan Hall, Jenny Hamilton, Kamran Khan, Gale Kittle, Marlee Lovett, Linda Nelson, Marissa Pabon, Diana Rede, Patrick Bolger, Ahmed Fetouh, Joan Woodcook, Cornelia Kamp, Julie Kennedy, Andrew McGarry, and Margherita Torti Data Availability Statement The data that support the findings of this study are available from the corresponding author upon reasonable request. References Associated Data Supplementary Materials Table S1: Baseline characteristics comparing participants included in versus excluded from the primary analysis set. Table S2: Correlation between at‐home FVC and in‐clinic SVC assessed < 14 days apart for the full study cohort and stratified by concurrent ALSFRS‐R bulbar domain score, grouped to achieve similar sample sizes. Table S3: Prevalence of in‐clinic SVC ≥ 60%‐predicted and operating characteristics using at‐home FVC ≥ 60%‐predicted to predict eligibility based on in‐clinic SVC ≥ 60%‐predicted. Table S4: Estimated progression rates over 48–52 weeks. Figure S1: (A) Scatterplot of at‐home FVC estimates versus in‐clinic SVC. Dashed lines indicate threshold for hypothetical eligibility criterion ≥ 60%‐predicted. (B) Receiver‐operator characteristic (ROC) curve for prediction of in‐clinic VC ≥ 60%‐predicted by concurrent at‐home VC. C. Positive predictive value (PPV) versus negative predictive value (NPV) for prediction of in‐clinic VC ≥ 60%‐predicted by concurrent at‐home VC. (D) Precision‐recall curve for prediction of in‐clinic VC ≥ 60%‐predicted by concurrent at‐home VC. Note that the y‐axis of panels C and D begin at 0.80. Data Availability Statement The data that support the findings of this study are available from the corresponding author upon reasonable request.