CNS Summit 2016 Abstracts of Poster Presentations

| February 1, 2017 | 0 Comments

Innov Clin Neurosci. 2017;14(1–2)3–23

A MESSAGE FROM THE EDITOR

Dear Colleagues:

Welcome to the annual CNS Summit Abstracts of Poster Presentations supplement to Innovations in Clinical Neuroscience. We are pleased to provide you with this reference guide to some of the innovative research that was presented during CNS Summit 2016, which was held October 27 to 30 in Boca Raton, Florida. The supplement is also available as a FLIPBOOK, a digital replica of the printed edition.

This supplement is just a small representation of what has become the premier event each year in drug development. Over 500 leaders in the field attended CNS Summit 2016 to experience more than 50 illuminating talks and witness the unveiling of 26 new innovations, such as Zephyr Illuminate™, a next generation enterprise solution that integrates global health data from thousands of disconnected sources, and AiCure’s artificial intelligence platform, a mobile technology that provides visual confirmation of drug ingestion by patients. And during the meeting, #CNSSummit was the second leading hashtag trending on Twitter—a testament to the excitement and enthusiasm the Summit generates among its participants.

The CNS Summit is committed to collaboration among all stakeholders involved in drug development, and we believe collaboration and data sharing with those involved in research across a wide variety of disease states will create more opportunity for developing better, safer, and more accessible drugs universally and within all areas of healthcare.

In this abstract supplement, we’ve organized and cross-referenced CNS?Summit 2016 poster abstracts into the following groups for your convenience and easy reference:

  • Artificial Intelligence/Machine Learning
  • Biomarkers and Imaging
  • Computer-based Tools
  • Digital Medicine
  • Investigative Drug Compounds and Therapies
  • Mobile Technology
  • Patient Assessment and Adherence
  • Placebo Response
  • Rater Assessment and Training
  • Trial Methodology/Study Protocol

You will also find an alphabetical index by author and poster title on pages 22 and 23 of this publication.

We hope you find the CNS Summit 2016 poster abstract supplement informative and that it provides a useful snapshot of the research that is presented during the CNS Summit each year. Make sure to mark your calendars for CNS?Summit 2017, which will be November 16 to 19 at the Boca Raton Resort in Florida. We’ll have more enlightening talks and innovative reveals than ever before, and you won’t want to miss the opportunity to learn about the latest technology and innovations in drug development and network with leaders in the field. Use promotion code ICNS2017 when registering for the CNS Summit, and you will receive $100 off your registration fee. Visit www.cnssummit.org for more information.

Sincerely
Amir Kalali, MD
Editor, Innovations in Clinical Neuroscience


TABLE OF CONTENTS

Artificial Intelligence/Machine Learning

  • Evaluating the use of an artificial intelligence platform on mobile devices to measure adherence in subjects with an acute exacerbation of schizophrenia
  • NetraMark: predicting placebo and drug response for pharma

Biomarkers and Imaging

  • Comparison of actigraphy endpoints for estimating nocturnal scratching duration in patients with atopic dermatitis
  • Development of neurophysiological-based biomarkers for neurodegenerative disease and psychiatric disorders using EEG
  • Sleep staging based on a reduced montage using two EOG channels

Computer-based Tools

  • [An] electronic (eCOA) prompted Yale Global Tic Severity Scale (YGTSS) with blinded internal scoring
  • [The] electronic self-report of the C-SSRS (eC-SSRS) places little burden on patients initially and even less with repeated administrations
  • [The] NetSCID: A validated web-based adaptive version of the SCID that improves diagnostic accuracy in mental health research

Digital Medicine

  • Clinical trial of Clickotine®, a Digital therapeutics™ solution for smoking cessation: design and total health execution methodology
  • Development of neurophysiological-based biomarkers for neurodegenerative disease and psychiatric disorders using EEG
  • Evaluating the use of an artificial intelligence platform on mobile devices to measure adherence in subjects with an acute exacerbation of schizophrenia
    Optimizing the mobile clinical trial: data collection through a web-based platform in telemedicine protocol

Investigative Drug Compounds and Therapies

  • [The] AtEase study: treatment of military-related PTSD
  • [The] clinical development of ITI-007 for the treatment of schizophrenia
  • D3: deuterium drug development
  • Drug repositioning for neurological disorders using a sustained release formulation of exenatide
  • [The] effect of cholesterol on positive symptoms of schizophrenia

Mobile Technology

  • Clinical trial of Clickotine®, a Digital Therapeutics™ solution for smoking cessation: design and total health execution methodology
  • Evaluating the use of an artificial intelligence platform on mobile devices to measure adherence in subjects with an acute exacerbation of schizophrenia
  • Optimizing the mobile clinical trial: data collection through a web-based platform in telemedicine protocol

Patient Assessment and Adherence

  • Are there reporting differences at screening on the Montgomery-Äsberg Depression Rating Scale (MADRS) between older and younger adults when using remote assessments?
  • Cultural adaptation of the Virtual Reality Functional Capacity Assessment Tool (VRFCAT) for use in the United Kingdom and Canada
  • Development of neurophysiological-based biomarkers for neurodegenerative disease and psychiatric disorders using EEG
  • Discrepancies between CGI-S score and PANSS item level scores—an exploratory analysis
  • [An] electronic (eCOA)-prompted Yale Global Tic Severity Scale (YGTSS) with blinded internal scoring
  • [The] electronic self-report of the C-SSRS (eC-SSRS) places little burden on patients initially and even less with repeated administrations
  • Evaluating the use of an artificial intelligence platform on mobile devices to measure adherence in subjects with an acute exacerbation of schizophrenia
  • NetraMark: predicting placebo and drug response for pharma
  • [The] NetSCID: a validated web-based adaptive version of the SCID that improves diagnostic accuracy in mental health research
  • Patients with psychiatric diagnoses indicate willingness to report suicidal ideation and behavior more honestly by self-report than in face-to-face interviews
  • Translation of the National Institutes of Health Stroke Scale (NIHSS) list of words: methodology and challenges

Placebo Response

  • Factors associated with response rate in inpatient schizophrenic participants
  • NetraMark: predicting placebo and drug response for pharma

Rater Assessment and Training 

  • Understanding rater preferences in services used to increase the reliability of clinical trials: a multi-national survey

Trial Methodology/Study Protocol

  • Could a paradigm shift in consumptive disorders research benefit patients and sponsors alike? A smoking cessation case study reinforces “the manner in which we look at things determines what we see!”
  • Discrepancies between CGI-S score and PANSS item level scores—an exploratory analysis
  • Evaluating the use of an artificial intelligence platform on mobile devices to measure adherence in subjects with an acute exacerbation of schizophrenia
  • Methodology and management of clinical trials with adaptive designs
  • NetraMark: predicting placebo and drug response for pharma
  • [A] reanalysis of the effects of protocol design and subject stipend on completion rates in Phase 1 studies in subjects with stable schizophrenia or schizoaffective disorder: Does money matter?
  • Toward a better understanding of informant contributions to schizophrenia trial quality: data from the encenicline cognitive impairment program
  • Understanding rater preferences in services used to increase the reliability of clinical trials: a multi-national survey

ARTIFICIAL INTELLIGENCE/MACHINE LEARNING

Evaluating the use of an artificial intelligence platform on mobile devices to measure adherence in subjects with an acute exacerbation of schizophrenia

Presenters: Shafner L1, McCue M2, Rubin A2, Dong X2, Hanson E2, Mahableshwarkar A2, Hanina A1, Macek T2

Affiliations: 1AiCure, New York, New York; 2Takeda Development Center Americas, Inc., Deerfield, Illinois

Objective: The need to minimize medication nonadherence is particularly important in central nervous system (CNS) clinical trials. An artificial intelligence (AI) platform was assessed in measuring and increasing medication adherence in subjects with schizophrenia in a Phase 2, randomized, double-blind study.

Design: Subjects in the TAK-063 study who were stable after three or more weeks of inpatient treatment and were discharged were followed up as outpatients for the remainder of the six-week period. Subjects were given devices with the AI application downloaded and asked to use the application for dosing administrations. The primary adherence measure for the study was based on scheduled pill counts; AI platform adherence data were tested for exploratory purposes.

Results: The AI platform was used by 26 subjects for 372 subject days; 744 adherence parameters were collected. Five subjects discontinued early (19.2%). For subjects who completed the trial, mean (standard deviation [SD]) cumulative adherence rates based on visual confirmation of drug ingestion (AI application) and on pill count were 82.5 percent and 99.7 percent, respectively. The mean time to use the AI platform was 86.8 seconds per pill.

Conclusion: Subjects with acutely exacerbated schizophrenia who were eligible for discharge from the inpatient setting and who completed the study demonstrated high rates of adherence using the mobile AI application. Subjects were able to easily use the technology. Use of the platform did not appear to increase the dropout rate. This study demonstrates the feasibility of using AI platforms to ensure high adherence, provide reliable adherence data, and rapidly detect nonadherence in CNS trials.

Disclosures/funding: Adam Hanina and Laura Shafner are employees of AiCure, New York, New York, and consultants to Takeda. Xinxin Dong, Elizabeth Hanson, Thomas A. Macek, Atul Mahableshwarkar, and Maggie McCue are employees of Takeda Development Center Americas, Inc., Deerfield, Illinois. Anne Rubin is a former employee of Takeda.

 

NetraMark: predicting placebo and drug response for pharma

Presenters: Geraci J

Affiliations: NetraMark Corp., Ontario, Canada

Objective: We wished to showcase precision medicine through novel machine learning models based on molecular and psychiatric scale data that can impact clinical trial success through placebo response and drug efficacy prediction.

Design: Utilizing data sets, we discovered models of placebo response robust enough to reproduce (within the specific treatment paradigm of a trial) and models that demonstrated the patient population purification process for predicting response in a precise way.

Results: We found two models of placebo response that had an accuracy of over 85 percent. For some situations we were able to predict if someone was a placebo responder with an accuracy of over 95 percent. Clinical scales alone were used for one of these models. We also based models on miRNA expression data that were capable of accurately predicting response for a particular subpopulation of major depression patients. This provided evidence that our technology is capable of dealing with heterogeneous patient populations. This could be valuable during Phase 3 scenarios for drug protocols that only work for a portion of the treatment population.

Conclusion: NetraMark developed a machine learning system to help pharmaceutical companies who must deal with complex patient populations. NetraMark accomplished this by providing accurate models of placebo and drug response. In situations where the efficacy of a drug is limited to a subpopulation whose effect could be “washed out” by the United States Food and Drug Adminisation (FDA) statistical paradigm, we provided what we call a Patient Population Purification process. This process is capable of predicting who these patients are with a high level of accuracy. Our unique system has the ability to help companies get potential medications on the market through our approach to precision medicine.

Disclosures/funding: We are now working with several pharmaceutical companies. At this time, we are unable to disclose all our current relationships.


BIOMARKERS AND IMAGING

Comparison of actigraphy endpoints for estimating nocturnal scratching duration in patients with atopic dermatitis

Presenters: Peterson B1, Moreau A1, Anderer P1, Cerny A1, Ross M1, Almazan T2, Craft N2

Affiliations: 1Philips Healthcare, Bend, Oregon; 2Science 37, Los Angeles, California.

Objective: Wrist-worn actigraphy measures of total sleep time (TST), wake after sleep onset (WASO), and sleep efficiency (Eff) have been used as indicators of nocturnal scratching. We compared those indicators with an algorithm that uses high-resolution actigraphy and neural networks (NN) to determine the duration of scratching events.

Design: Six healthy controls and 18 patients with atopic dermatitis (AD) wore actigraphy devices on each wrist and were video recorded during one night in a sleep lab. The videos were scored to determine the true duration of scratching (% time). TST, WASO, and Eff were calculated using standard actigraphy algorithms. A regression line of each endpoint against the true scratching duration was used to estimate the duration of scratching events from the sleep endpoints. A new NN algorithm calculated scratching duration directly from the actigraphy data. The error was the difference between the estimate and the true value.

Results: The true scratching duration (% time) was 0.1±0.1 standard deviation (SD) for healthy subjects and 3.6±6.2 for patients with AD (p<0.03). All endpoints correlated with the true scratching duration (r=0.97 [NN], 0.64 [TST], 0.70 [WASO], 0.69 [Eff]) but NN had the best correlation and the lowest median error (0.4% time vs. 1.6% time [TST], 1.5% time [WASO], and 1.3% time [Eff]).

Conclusion: Compared to the use of conventional sleep endpoints, the direct measurement of nocturnal scratching duration with the NN algorithm provided the strongest correlation with the true value and the smallest error.

Disclosures/funding: None reported.

 

Development of neurophysiological-based biomarkers for neurodegenerative disease and psychiatric disorders using EEG

Presenters: Waninger S1, Berka C1, Stikic M1, Korszen S1, Salat D2, Verma A3

Affiliations: 1Advanced Brain Monitoring, Inc., Carlsbad, California; 2MGH/MIT/HMS Athinoula A. Martinos Center for Biomedical Imaging, Department of Radiology, Massachusetts General Hospital, Charlestown, Massachusetts ; 3Biogen, Cambridge, Massachusetts

Background: Successful drug development for neurodegenerative diseases and psychiatric disorders requires objective, reliable, and accurate measures to evaluate disease progression and therapeutic efficacy. The potential of neurophysiological biomarkers using electroencephalography (EEG) has been highlighted in ongoing studies at Advanced Brain Monitoring on Parkinson’s disease (PD), Alzheimer’s disease (AD), mild cognitive impairment (MCI), mood disorders, and posttraumatic stress disorder (PTSD).

Design: Using our wireless B-Alert X24 system with a standard 10–20 montage, we acquired and analyzed EEG data both during resting state and neurocognitive tasks designed to activate the neural circuits involved in attention, memory, and emotion and elicit event-related potentials (ERPs). For resting state, data were acquired for five minutes with eyes closed (EC) and five minutes with eyes open (EO). These data were decontaminated and converted from the time domain to the frequency domain using Fast Fourier Transform (FFT) to calculate power spectral densities (PSD) grouped into the standard EEG bandwidths (delta, theta, alpha, beta, and gamma).

Results: Comparison of PSD from AD and MCI patients to healthy, age-matched controls indicated distinguishing features, particularly the “slowing” of EEG exemplified by an increase in slow wave bands and a decrease in fast power that is typically observed in patients with cognitive decline. The source of the abnormal EEG in MCI patients was localized to the middle and superior temporal gyrus and fusiform gyrus using low resolution electromagnetic tomography (LORETA). Variables extracted from the resting state data (i.e., absolute PSDs, relative PSDs, and wavelets) are grouped together into a feature vector and the most discriminative variables are selected to construct linear discriminate function analysis (lDFA) models. Application of lDFA to the MCI dataset results in high accuracy and specificity using auto-validation or leave-one-out cross-validation. Although a great deal of information is derived from resting state data, we also have the capability to capture engagement of neural circuits during neurocognitive tasks that stimulate and elicit neural patterns associated with attention, memory and emotion. One such task, three-choice vigilance (3CVT), is a visual choice reaction time task designed to measure sustained attention and target detection. EEG data acquired during the 3-CVT task indicates significantly longer peak latency in the parietal N2 component of the ERP in a PTSD cohort compared to healthy control cohort.

Conclusion: These data provide further evidence that PTSD patients have increased difficulty with attentional resources and identifies a potential biomarker for PTSD disease progression. Both resting state and event related EEG data have potential for use as cost efficient, noninvasive pharmacodynamic endpoints in neurodegenerative disease and psychiatric disorder clinical trials of experimental therapeutics.

Disclosures/funding: None reported.

 

NetraMark: predicting placebo and drug response for pharma

Presenters: Geraci J

Affiliations: NetraMark Corp., Ontario, Canada

Objective: We wished to showcase precision medicine through novel machine learning models based on molecular and psychiatric scale data that can impact clinical trial success through placebo response and drug efficacy prediction.

Design: Utilizing data sets, we discovered models of placebo response robust enough to reproduce (within the specific treatment paradigm of a trial) and models that demonstrated the patient population purification process for predicting response in a precise way.

Results: We found two models of placebo response that had an accuracy of over 85 percent. For some situations we were able to predict if someone was a placebo responder with an accuracy of over 95 percent. Clinical scales alone were used for one of these models. We also based models on miRNA expression data that were capable of accurately predicting response for a particular subpopulation of major depression patients. This provided evidence that our technology is capable of dealing with heterogeneous patient populations. This could be valuable during Phase 3 scenarios for drug protocols that only work for a portion of the treatment population.

Conclusion: NetraMark developed a machine learning system to help pharmaceutical companies who must deal with complex patient populations. NetraMark accomplished this by providing accurate models of placebo and drug response. In situations where the efficacy of a drug is limited to a subpopulation whose effect could be “washed out” by the United States Food and Drug Administration (FDA) statistical paradigm, we provided what we call a Patient Population Purification process. This process is capable of predicting who these patients are with a high level of accuracy. Our unique system has the ability to help companies get potential medications on the market through our approach to precision medicine.

Disclosures/funding: We are now working with several pharmaceutical companies. At this time, we are unable to disclose all our current relationships.

 

Sleep staging based on a reduced montage using two EOG channels

Presenters: Dorffner G

Affiliations: The Siesta Group, Wien, Austria; Medical University of Vienna, Section for Artificial Intelligence and Decision Support, Vienna, Austria

Objective: Currently, objective assessment of sleep architecture and sleep continuity in clinical trials relies on the recording of distinct biological signals (e.g., electroencephalography [EEG], electrooculography [EOG], and electromyography [EMG]) for a full night. This method—polysomnography (PSG)—is usually performed at specialized sleep labs requiring skilled personnel and full equipment, which, being expensive and to some extent burdensome for the patient, limits the number measurements to only a few nights in a protocol. This might not be representative for a patient’s sleep. Thus, a portable, less intrusive, and self-applicable solution for sleep measurement would allow for the acquisition of more nights in the patient’s familiar environment. The aim of this study was to investigate if a reduced setting requiring two EOG channels only would yield comparable results to a full PSG that includes six EEGs, two EOGs, and one EMG channel.

Design: Sleep recordings from 36 healthy controls (2 nights each) were analyzed using a validated computer-assisted scoring system. Only the standard two EOG channels were used as input data, which were submitted to a modified version of the analyzer.

Results: The main three states wakefulness (r=0.87), non-rapid eye movement (NREM) sleep (r=0.77), and rapid eye movement (REM) sleep (r=0.68), were identified effectively (full montage vs. reduced montage). On an epoch-by-epoch basis, Cohens Kappa was 0.65 (“good agreement”) with agreement rates of 86 percent for waking (W), 81 percent for slow wave sleep (SWS), and 84 percent for REM.

Conclusion: This work provides promising evidence that, with the proper modification of existing computer-based sleep scoring solutions, a reduced montage permits sleep measurements that lead to results comparable to full PSG, at least with respect to many important sleep variables.

Disclosures/funding: Georg Dorffner is employee and shareholder of The Siesta Group, Wein, Austria.


COMPUTER-BASED TOOLS

An electronic (eCOA)-prompted Yale Global Tic Severity Scale (YGTSS) with blinded internal scoring

Presenters: Busner J1, Farber R2, O’Brien C2, Liang G2, Scahill L3, Coffey B4

Affiliations: 1Penn State College of Medicine, Department of Psychiatry, Hershey, Pennsylvania, and Bracket Global, Wayne, Pennsylvania;2 Neurocrine Biosciences, Inc., San Diego, California; 3Emory University, Atlanta, Georgia; 4Icahn School of Medicine at Mount Sinai School, New York, New York

Objective: The YGTSS is a gold-standard tic severity assessment in Tourette’s syndrome (TS) studies. To improve ratings quality, we developed an eCOA-prompted YGTSS that provided rating guidance, captured rater scores, and generated rater-blinded algorithm-derived scores as quality checks.

Design: An eCOA-prompted YGTSS was developed with TS experts. The eCOA YGTSS ensured correct navigation through the scale and assisted raters by displaying earlier endorsed tics, when needed, for motor and phonic severity ratings. The scale included algorithms for a second set of rater-blinded “tandem” scores. The scale is being piloted in two ongoing, placebo-controlled, multisite TS trials, one pediatric and one adult, with rater scores serving as efficacy data.

Results: There were 99 visits by 20 raters that were completed at the time of the analysis.  Correlations between rater and computer scores were high for each of the 10 YGTSS severity scores (range: 0.74–0.91, all p values <0.0001); for the Total Tic Score (TTS) (primary efficacy measure), the correlation was 0.95 (p<0.0001).  The mean rater versus computer TTS scores were almost identical (28.8 and 28.5, respectively, NS). The findings did not differ between pediatric and adult subjects.

Conclusion: Our internal scoring algorithms correlated significantly with all rater-selected motor, phonic, and TTS scores, with the latter nearly identical. The work provides preliminary validation of our algorithms and supports the feasibility of the approach. In a risk-based monitoring model, less trained raters whose scores deviate significantly from those of the internal algorithm might be selected for additional scrutiny and intervention.

Disclosures/funding: Dr. Busner is an employee of Bracket Global, Wayne, Pennsylvania. Drs. Farber, O’Brien, and Liang are employees of Neurocrine Biosciences, Inc., San Diego, California. Dr. Scahill is a consultant to Roche, Bracket, Neuren Pharmaceuticals, MedAdvante, Coronado Biosciences, Inc., and Supernus Pharmaceuticals, Inc.; has received research support from the United States National Institute of Mental Health (NIMH) and Department of Defense; and has received royalties from Oxford and Guilford. Dr. Coffey has received honoraria from American Academy of Child and Adolescent Psychiatry; has received research support from Neurocrine, NIMH, Shire, and Catalyst Pharmaceuticals Inc.; serves on the advisory board, received Center of Excellence Funding, and is on the speakers bureau of Tourette Association of America; serves on the advisory board and receives research support from Auspex Pharmaceuticals; and serves on the advisory board of Genco Sciences.

 

The electronic self-report of the C-SSRS (eC-SSRS) places little burden on patients initially and even less with repeated administrations

Presenters: Yamamoto R1, Durand E1, Lima V1, Christopher D1, Yershova K2, Dallabrida S1

Affiliations: 1ERT Corp., New York, New York2; Columbia University/New York State Psychiatric Institute, New York, New York

Objective: A key factor to consider for instruments that measure suicidal ideation and behavior (SIB) is the duration of patient completion time. The Columbia Suicide Severity Rating Scale (C-SSRS) (interviewer completed) and electronic C-SSRS (eC-SSRS) (self-report) are sourced by the United States Food and Drug Administration (FDA) in the SIB Industry Guidance as recommended assessments. This study examined patient completion time for the tablet-based eC-SSRS over time.

Design: Over 2,000 subjects with substance abuse disorder completed the tablet-based eC-SSRS assessment at each clinic visit in a trial. Data from the eC-SSRS since last contact version were analyzed.

Results: Completion times for patients who scored negative for SIB using the tablet version of the e-C-SSRS were 1.15±1.33min at Visit 1, 0.84±0.89min at Visit 7, and 0.73±0.79min at Visit 12. There was a significant decrease in the time it took patients to complete the e-CSSRS across Visits 1 through 12 General Linear Model (mixed) (F=5809.25, p<0.0001). ANOVA repeated measures on all 12 visits (F=7.7, p<0.0001) showed that the time to complete the eC-SSRS decreased across visits. A t-test between Visit 1 and Visit 12 (p<0.0001, t=4.8) indicated that it took significantly less time for subjects to complete the eC-SSRS on Visit 12 versus Visit 1.

Conclusion: These data showed that there was little burden on subjects to complete the tablet eC-SSRS, and this burden decreased with repeated administrations. This is a particularly important consideration in lengthy studies with many repeat collections of an SIB assessment.

Disclosures/funding: Drs. Yamamoto, Durand, and Dallabrida and Ms. Lima and Mr. Christopher are employees of ERT Corp., New York, New York. Dr. Yershova is an employee of Columbia University/New York State Psychiatric Institute, New York, New York.

 

The NetSCID: a validated web-based adaptive version of the SCID that improves diagnostic accuracy in mental health research

Presenters: Brodey BB1, Zweede L1

Affiliations: 1TeleSage, Inc., Chapel Hill, North Carolina

Background: The Structured Clinical Interview for the Diagnostic and Statistical Manual of Mental Disorders (SCID, DSM) is the gold standard for research-based mental health diagnoses. It is the most widely used comprehensive tool for assessing DSM diagnoses. Its direct adherence to DSM criteria provides strong test-retest and high interrater reliability; however, administration of the full research version averages two hours and requires considerable clinician training, making it impractical for many protocols.

Objective: Our objective was to develop and validate a highly configurable and secure web-based version of the SCID—the NetSCID— thereby making the gold standard in mental health diagnostics available for clinical trials.
Methods: The validation research included 24 clinicians who administered the SCID to 230 participants that completed the paper SCID and/or the NetSCID. Data-entry errors, branching errors, and clinician satisfaction were quantified.

Results: Ninety-seven percent of error rates occurred among clinicians administering the paper SCID; all errors that occurred when utilizing the NetSCID were subsequently corrected by our programmers. Administration time was reduced by over 30 percent. Clinicians found it easier to administer (p< 0.05), easier to navigate (p< 0.05), and simpler to score (p< 0.01). Ninety percent of clinicians preferred the NetSCID to its paper counterpart.

Conclusion: Because of its unique ability to record symptoms in a standardized database, the NetSCID facilitates characterization of patient populations and assists with identification of sub-groups that may respond to interventions. Adoption of this tool has shown to improve diagnostic accuracy and will increase the power in central nervous system (CNS) clinical trials worldwide.

Disclosures/funding: This research was supported in part by a grant from the National Institutes of Mental Health, which was awarded to TeleSage, Inc. Dr. Brodey is Chief Executive Officer of TeleSage, Inc., and Lisa Zweede is employed by TeleSage, Inc. as a clinical research specialist.


DIGITAL MEDICINE

Clinical trial of Clickotine®, a Digital Therapeutics™ solution for smoking cessation: design and total health execution methodology

Presenters: Steinerman J1, Klein D1, Silver T1, Berger A1, Luo S1, Schork N1

Affiliations: 1 Click Therapeutics, New York, New York

Objective: The objective of this review was to describe the efficient and effective execution of a clinical trial of a digital therapeutic for smoking cessation.

Design: Clickotine is a mobile health solution for smoking cessation that combines evidence-based behavioral interventions with a unique engagement strategy that leverages personalization and contextualization. The potential utility of Clickotine was assessed in an open-label, eight-week study that focused on user-engagement, safety, tolerability, and smoking behavior efficacy. Key entry criteria included being 18 to 65 years of age, smoking five or more cigarettes per day, an interest in quitting smoking, ownership of an iPhone, living in the United States, and providing informed consent. The study employed remote telehealth interactions only. Designated sponsor representatives interacted directly with participants, while medical and scientific team members monitored and analyzed de-identified data. Participants were recruited via social media and a pre-screening telephone call. Informed consent and study questionnaires were delivered via an internet portal, complementing in-app data capture. Procedures for medical monitoring and biochemical verification of smoking cessation were in place throughout the study, which was institutional review board approved.

Results: After 63 days of social media recruitment, 2,050 contacts were received and 617 telephone pre-screens were conducted, resulting in 452 participants providing online informed consent. Participants totaling 416 ultimately made up an intention-to-treat population broadly representative of United States smokers who want to quit. Baseline characteristics and details of telehealth trial execution were presented.

Conclusion: Digital health enterprises can execute high-quality clinical trials by implementing social media recruitment, digital data capture, and remote interactions, while sparing resources and prioritizing participant convenience.

Disclosures/funding: This study was funded by Click Therapeutics, New York, New York. JS is a consultant and shareholder of Click Therapeutics and an employee and shareholder of Teva Pharmaceuticals, Malvern, Pennsylvania. DK is an officer and shareholder of Click Therapeutics. TS is an officer and shareholder of Click Therapeutics. AB is an employee and shareholder of Click Therapeutics. SL is a consultant of Click Therapeutics. NS is an officer and shareholder of Click Therapeutics.

 

Development of neurophysiological-based biomarkers for neurodegenerative disease and psychiatric disorders using EEG

Presenters: Waninger S1, Berka C1, Stikic M1, Korszen S1, Salat D2, Verma A3

Affiliations: 1Advanced Brain Monitoring, Inc., Carlsbad, California; 2MGH/MIT/HMS Athinoula A. Martinos Center for Biomedical Imaging, Department of Radiology, Massachusetts General Hospital, Charlestown, Massachusetts ; 3Biogen, Cambridge, Massachusetts

Background: Successful drug development for neurodegenerative diseases and psychiatric disorders requires objective, reliable, and accurate measures to evaluate disease progression and therapeutic efficacy. The potential of neurophysiological biomarkers using electroencephalography (EEG) has been highlighted in ongoing studies at Advanced Brain Monitoring on Parkinson’s disease (PD), Alzheimer’s disease (AD), mild cognitive impairment (MCI), mood disorders, and posttraumatic stress disorder (PTSD).

Design: Using our wireless B-Alert X24 system with a standard 10–20 montage, we acquired and analyzed EEG data both during resting state and neurocognitive tasks designed to activate the neural circuits involved in attention, memory, and emotion and elicit event-related potentials (ERPs). For resting state, data were acquired for five minutes with eyes closed (EC) and five minutes with eyes open (EO). These data were decontaminated and converted from the time domain to the frequency domain using Fast Fourier Transform (FFT) to calculate power spectral densities (PSD) grouped into the standard EEG bandwidths (delta, theta, alpha, beta, and gamma).

Results: Comparison of PSD from AD and MCI patients to healthy, age-matched controls indicated distinguishing features, particularly the “slowing” of EEG exemplified by an increase in slow wave bands and a decrease in fast power that is typically observed in patients with cognitive decline. The source of the abnormal EEG in MCI patients was localized to the middle and superior temporal gyrus and fusiform gyrus using low resolution electromagnetic tomography (LORETA). Variables extracted from the resting state data (i.e., absolute PSDs, relative PSDs, and wavelets) are grouped together into a feature vector and the most discriminative variables are selected to construct linear discriminate function analysis (lDFA) models. Application of lDFA to the MCI dataset results in high accuracy and specificity using auto-validation or leave-one-out cross-validation. Although a great deal of information is derived from resting state data, we also have the capability to capture engagement of neural circuits during neurocognitive tasks that stimulate and elicit neural patterns associated with attention, memory and emotion. One such task, three-choice vigilance (3CVT), is a visual choice reaction time task designed to measure sustained attention and target detection. EEG data acquired during the 3-CVT task indicates significantly longer peak latency in the parietal N2 component of the ERP in a PTSD cohort compared to healthy control cohort.

Conclusion: These data provide further evidence that PTSD patients have increased difficulty with attentional resources and identifies a potential biomarker for PTSD disease progression. Both resting state and event related EEG data have potential for use as cost efficient, noninvasive pharmacodynamic endpoints in neurodegenerative disease and psychiatric disorder clinical trials of experimental therapeutics.

Disclosures/funding: None reported.

 

Evaluating the use of an artificial intelligence platform on mobile devices to measure adherence in subjects with an acute exacerbation of schizophrenia

Presenters: Shafner L1, McCue M2, Rubin A2, Dong X2, Hanson E2, Mahableshwarkar A2, Hanina A1, Macek T2

Affiliations: 1AiCure, New York, New York; 2Takeda Development Center Americas, Inc., Deerfield, Illinois
Objective: The need to minimize medication nonadherence is particularly important in central nervous system (CNS) clinical trials. An artificial intelligence (AI) platform was assessed in measuring and increasing medication adherence in subjects with schizophrenia in a Phase 2, randomized, double-blind study.

Design: Subjects in the TAK-063 study who were stable after three or more weeks of inpatient treatment and were discharged were followed up as outpatients for the remainder of the six-week period. Subjects were given devices with the AI application downloaded and asked to use the application for dosing administrations. The primary adherence measure for the study was based on scheduled pill counts; AI platform adherence data were tested for exploratory purposes.

Results: The AI platform was used by 26 subjects for 372 subject days; 744 adherence parameters were collected. Five subjects discontinued early (19.2%). For subjects who completed the trial, mean (standard deviation [SD]) cumulative adherence rates based on visual confirmation of drug ingestion (AI application) and on pill count were 82.5 percent and 99.7 percent, respectively. The mean time to use the AI platform was 86.8 seconds per pill.

Conclusion: Subjects with acutely exacerbated schizophrenia who were eligible for discharge from the inpatient setting and who completed the study demonstrated high rates of adherence using the mobile AI application. Subjects were able to easily use the technology. Use of the platform did not appear to increase the dropout rate. This study demonstrates the feasibility of using AI platforms to ensure high adherence, provide reliable adherence data, and rapidly detect nonadherence in CNS trials.

Disclosures/funding: Adam Hanina and Laura Shafner are employees of AiCure, New York, New York, and consultants to Takeda. Xinxin Dong, Elizabeth Hanson, Thomas A. Macek, Atul Mahableshwarkar, and Maggie McCue are employees of Takeda Development Center Americas, Inc., Deerfield, Illinois. Anne Rubin is a former employee of Takeda.

 

Optimizing the mobile clinical trial: data collection through a web-based platform in telemedicine protocol

Presenters: Shaw M1, Song G2, Charvet L1, Beringer J3, Waibel U3

Affiliations: 1NYU School of Medicine, New York, New York; 2Stony Brook Medicine, Stony Brook, New York; 3Berisoft Corp., Redwood City, California

Objective: A web-based platform was employed to administer cognitive testing and self-reported mood in a clinical trial where the intervention was administered remotely to participants at home through telerehabilitation.

Design: Using a remotely supervised telemedicine protocol, our study tested the benefit of transcranial direct current stimulation (tDCS) for participants with multiple sclerosis (MS). Participants completed daily assessments along with their treatments through a study-provided laptop. To avoid paperwork and to access data in real-time, we utilized the ERTSLab online platform, administrable from any computer with an internet connection. Cognitive functioning was measured by the Attention Networks Test-Interaction (ANT-I), and self-reported mood was measured using the Positive and Negative Affect Schedule (PANAS) taken from the Cognition Library of ERTSLab.

Results: For the ANT-I, excellent form reliability was found between ERTSLab and the PC-installed ePrime version. Form reliability was excellent (r=0.90). Using the ERTSLab version, in our MS sample we found significant improvement in alerting scores for the treatment (n=6) versus control (n=4) conditions (p=0.031). For the PANAS self-reported mood measure, online versus standard forms were similarly validated (r=0.90). For the study, 13 participants completed a daily PANAS assessment along with their treatment sessions, and we found a trend toward statistically significant improvement in positive (p=0.001) and negative (p=0.70) affects for those in the treatment versus control condition.

Conclusion: Online platforms, such as ERTSLab ,provide a mobile platform to administer cognitive assessment and self-reported symptom inventories in study designs where participants are assessed in remote locations including the use of telerehabilitation.

Disclosures/funding: None reported.


INVESTIGATIVE DRUG COMPOUNDS AND THERAPIES

The AtEase study: treatment of military-related PTSD

Presenters: Peters P1, Gendreau R2, Gendreau J1, Sullivan G1, Peters A1, Engels J3, Schaberg A4, Jividen H1, Lederman S1

Affiliations: 1Tonix Pharmaceuticals, Inc., New York, New York; 2Gendreau Consulting, Poway, California; 3Engels Statistical Consulting LLC, Minneapolis, Minnesota; 4Schaberg Consulting, Cary, North Carolina

Objective: Our objective was to perform a post-hoc analysis of the safety and efficacy of TNX-102 SL 5.6mg compared to placebo in posttraumatic stress disorder (PTSD) patients, as assessed by the Clinician-Administered PTSD Scale for DSM-5 (CAPS-5) score.

Design: ‘AtEase’ was a Phase 2, multicenter, 12-week, randomized, controlled trial in adults meeting a Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (DSM-5) diagnosis of PTSD as assessed by CAPS-5. Patients were randomized to TNX-102 SL 2.8mg, 5.6mg, or placebo in a 2:1:2 ratio. Eligible participants had PTSD criterion “A” qualifying trauma(s) during military service since 2001 with a baseline CAPS-5 severity score of 29 or greater. The primary efficacy endpoint was mean change from baseline (MCFB) in total CAPS-5 score at Week 12.

Results: Compared to placebo (n=92), patients who received TNX-102 SL 5.6mg (n=49) experienced a substantially greater MCFB in CAPS-5 at Week 12 (p=0.053; effect size (ES)=0.36). A post-hoc analysis of the subset with a baseline CAPS-5 score of 33 or higher which was shown to be a closer estimate to prior CAPS version entry scores of 50 or higher (used in previous studies), demonstrated larger separation from placebo. Treatment differences between placebo (n=77) and TNX-102 SL 5.6mg (n=38) were significant for the primary endpoint of CAPS-5 (p=0.013; ES=0.53) as well as on multiple secondary endpoints. The most commonly reported adverse events in this subset were oral hypoaesthesia, somnolence, and dry mouth, similar to the original study population.

Conclusion: TNX-102 SL 5.6mg was superior to placebo with a substantial ES when adjusting the CAPS-5 minimum entrance criteria to match previous PTSD studies.

Disclosures/funding: TNX-102 SL is an investigational new drug and has not been approved for any indications.

 

The clinical development of ITI-007 for the treatment of schizophrenia

Presenters: Vanover K, Glass S, O’Gorman C, Saillar J, Weingart M, Correll C, Mates S, Davis R

Affiliations: Intra-Cellular Therapies, Inc., New York, New York

Objective: ITI-007, an investigational agent acting through serotonergic, dopaminergic, and glutamatergic systems, is in development for schizophrenia, bipolar depression, and agitation associated with dementia. Here we report on the schizophrenia program.

Design: ITI-007-005 was a four-week Phase 2 trial, wherein 335 patients were randomized to receive orally once-daily: ITI-007 (60mg or 120mg), risperidone (positive control), or placebo. ITI-007-301 was the first of two Phase 3 trials, wherein 450 patients were randomized to receive either ITI-007 (60mg or 40mg) or placebo for four weeks. ITI-007-302 was the second Phase 3 trial, and 696 patients were randomized to receive ITI-007 (60mg or 20mg), risperidone (positive control), or placebo for six weeks. The primary efficacy endpoint is Positive and Negative Symptom Scale (PANSS) total score change from baseline versus placebo.

Results: In ITI-007-005 and ITI-007-301, ITI-007 60mg met the primary endpoint. In ITI-007-301, ITI-007 60mg, with no dose titration, showed early and sustained efficacy on both the PANSS total and Positive Symptom Subscale and met the key secondary endpoint with significant improvement on CGI-S. In ITI-007-301, ITI-007 40mg separated significantly from placebo on the PANSS Positive Symptom subscale as well as CGI-S. Pro-social benefits were also observed. ITI-007 was safe and well-tolerated with a motor and cardio-metabolic safety profile similar to placebo. The ITI-007-302 trial is ongoing, and results will be presented, if available.

Conclusion: ITI-007 represents a new approach to the treatment of schizophrenia with unique pharmacologic properties and a differentiating clinical profile.

Disclosures/funding: KEV, SJG, COG, JS, MW, SM, RED are full-time employees of Intra-Cellular Therapies. CUC financial disclosures include Alkermes, Bristol-Myers Squibb, Forum, Gerson Lehrman Group, Intra-Cellular Therapies, Janssen/J&J, Lundbeck, Medavante, Medscape, Otsuka, Pfizer, ProPhase, Sunovion, Supernus, Takeda, and Teva.

 

D3: deuterium drug development

Presenters: Barag J, Laage T, Leathers T

Affiliations: Premier Research, Trenton, New Jersey

Objective: The objective of this review was to highlight recent developments and challenges in using deuterium atom substitution to create novel chemical entities that prolong the metabolism and therefore increase the therapeutic effect of parent compounds. Deuterium is an isotope of hydrogen having double the atomic weight (1–2) because of the addition of a neutron to the single proton in the nucleus of the most common isotope.

Design: We reviewed the physiochemical (or pharmacodynamic) effects of such substitutions on the action of parent compounds (usually minimal) and the contrast with the metabolic (or pharmacokinetic) consequences of deuteration, as the strength of the covalent carbon-deuterium bond is 10 times that of the carbon-hydrogen bond (kinetic isotope effect), creating more resistance to chemical or enzymatic cleavage. If the cleavage of this bond is a metabolic, rate-determining step (often the case where oxidation or hydroxylation occurs), favorable effects of the parent compound are prolonged and deleterious effects of metabolites reduced. Intellectual property challenges involving patentability (including “art” and “obviousness”) were also reviewed. Deuterated compounds currently in development (e.g., Auspex—tetrabenazine and pirfenidone; Sage—allopregnanolone; Amorsa—norketamine; Concert—dextromethorphan, ivacaftor, sodium oxybate, apremilast, ruxolitinib) and discontinued (Merck’s fludalanine) were also reviewed.

Conclusion: Deuteration is a powerful technique for developing and marketing versions of existing drugs with improved half-lives and reduced toxicity.

Disclosures/funding: None reported.

 

Drug repositioning for neurological disorders using a sustained release formulation of exenatide

Presenters: Kim D1,2, Li Y1,Tamargo I1, Kim H2, Tweedie D1, Wang Y3, Lee H2, Han B4, Greig N1

Affiliations: 1National Institute on Aging, Baltimore, Maryland; 2Peptron Inc., Daejeon, Korea; 3National Health Research Institutes, Taiwan; 4KRIBB, Daejeon, Korea.

Background: SmartDepotTM is Peptron’s proprietary technology for the sustained release (SR) microsphere formulation of various agents to provide once weekly or longer dosing regimens. exenatide, a long-acting GLP-1 analogue that is efficacious in type 2 diabetes mellitus (T2DM), has been identified as a novel treatment strategy for neurodegenerative disorders, including chronic diseases, such as Alzheimer’s disease (AD) and Parkinson’s disease (PD), and acute ones, such as traumatic brain injury (TBI). Consequent to its neuroprotective and neurotrophic actions, clinical trials in several of these disorders are on going. In prior preclinical research that supported these clinical studies, exenatide proved particularly efficacious when delivered continuously by a micro-osmotic pump in PD and mild TBI animal model studies. The application of the latest sustained release (SR) technology to exenatide, whereby steady-state concentrations can be effectively maintained over weeks to months, holds the potential to optimize the beneficial potential of a new drug treatment for a chronic disorder, particularly where disease-associated cognitive impairments may impact adherence.

Design and results: In this study, an exenatide SR-formulation (SR-exenatide) was evaluated in a concussive mild TBI mouse model and in a PD animal model. TBI and PD animals displayed deficits in behavioral tests associated with cognitive and motor performance, respectively, whereas SR-exenatide-treated animals performed similar to sham controls when treated with a clinically translatable dose used in ongoing T2DM clinical trials of SR-exenatide.

Conclusion: These behavioral data suggest a strong beneficial action of SR-exenatide in TBI and PD and are defining parameters for first human trials of SR-exenatide in these disorders. This convenient dosing regimen of SR-exenatide has applications for other diseases, such as AD, multiple system atrophy, and multiple sclerosis, where preclinical studies, likewise, have demonstrated promising exenatide actions.

 

The effect of cholesterol on positive symptoms of schizophrenia

Presenters: Tireman E, Templeton K, Kakar R, DeVito L, Pierre L, Pina D

Affiliations: Segal Institute for Clinical Research, Ft. Lauderdale, Florida

Background: Cholesterol may be a factor in the pathological development of schizophrenia, as it is an important molecular component of myelin and synapses within the brain. Currently, research has failed to investigate the effect of cholesterol on schizophrenia and instead has focused on other areas of mental illness. A relationship between cholesterol level and positive symptoms could allow physicians to increase the efficacy of antipsychotics by focusing treatment on patients with normal (neither high nor low) cholesterol. Additionally, future drug augmentation may look for agents that work on the cholesterol pathway to increase efficacy of decreasing positive symptoms. We examined positive symptoms of schizophrenia, as measured by the Brief Psychotic Rating Scale (BPRS), and fasting lipid profile to determine if lower levels of cholesterol are indicative of increased positive symptoms.

Design: Lipid profiles and BPRS scores were collected from 231 patients with schizophrenia who experienced an acute exacerbation of symptoms. Positive BRPS scores and BPRS total score at screening (i.e., items BPRS4 conceptual disorganization, BPRS10 hostility, BPRS11 suspiciousness, BPRS12 hallucinatory behavior, and BPRS15 unusual thought content) and fasting lipid profiles were analyzed to assess correlation.

Results: Correlational analyses were used to examine the relationship between fasting lipid profiles and BPRS scores. Results indicated an inverse relationship between the low density lipoprotein (LDL) and BPRS total score (r= -0.169, p=0.01). This suggests individuals with low LDL levels have higher BPRS total scores. Results also indicated an inverse relationship between the total cholesterol and BPRS total score (r= -0.133, p=0.044). This suggests individuals with low total cholesterol levels have higher BPRS scores at screening. Correlations between lipid profiles and individual positive items were non-significant.

Conclusion: The results of this analysis indicate that cholesterol levels are negatively related to BPRS total scores at screening. This association may be clinically substantial; however, it does not prove causality. It has been suggested that a reduced level of serum total cholesterol and LDL may be associated with increased aggression and overall symptoms of schizophrenia. These results may be relevant to clinical trials in that while it is standard to screen for high cholesterol, there could be a potential benefit in analyzing the LDL levels to assess the relationship to intrusive symptoms. Lipid profiles could potentially become an important biomarker to consider in the future.

Disclosures/funding: None reported.


MOBILE TECHNOLOGY

Clinical trial of Clickotine®, a Digital Therapeutics™ solution for smoking cessation: design and total health execution methodology

Presenters: Steinerman J1, Klein D1, Silver T1, Berger A1, Luo S1, Schork N1

Affiliations: 1 Click Therapeutics, New York, New York

Objective: The objective of this review was to describe the efficient and effective execution of a clinical trial of a digital therapeutic for smoking cessation.

Design: Clickotine is a mobile health solution for smoking cessation that combines evidence-based behavioral interventions with a unique engagement strategy that leverages personalization and contextualization. The potential utility of Clickotine was assessed in an open-label, eight-week study that focused on user-engagement, safety, tolerability, and smoking behavior efficacy. Key entry criteria included age 18 to 65 years, smoking five or more cigarettes per day, an interest in quitting smoking, ownership of an iPhone, living in the United States, and providing informed consent. The study employed remote telehealth interactions only. Designated sponsor representatives interacted directly with participants, while medical and scientific team members monitored and analyzed de-identified data. Participants were recruited via social media and a pre-screening telephone call. Informed consent and study questionnaires were delivered via an internet portal, complementing in-app data capture. Procedures for medical monitoring and biochemical verification of smoking cessation were in place throughout the study, which was institutional review board approved.

Results: After 63 days of social media recruitment, 2,050 contacts were received and 617 telephone pre-screens were conducted, resulting in 452 participants providing online informed consent. Participants totaling 416 ultimately made up an intention-to-treat population broadly representative of United States smokers who want to quit. Baseline characteristics and details of telehealth trial execution were presented.

Conclusion: Digital health enterprises can execute high-quality clinical trials by implementing social media recruitment, digital data capture, and remote interactions, while sparing resources and prioritizing participant convenience.

Disclosures/funding: This study was funded by Click Therapeutics, New York, New York. JS is a consultant and shareholder of Click Therapeutics and an employee and shareholder of Teva Pharmaceuticals, Malvern, Pennsylvania. DK is an officer and shareholder of Click Therapeutics. TS is an officer and shareholder of Click Therapeutics. AB is an employee and shareholder of Click Therapeutics. SL is a consultant of Click Therapeutics. NS is an officer and shareholder of Click Therapeutics.

 

Evaluating the use of an artificial intelligence platform on mobile devices to measure adherence in subjects with an acute exacerbation of schizophrenia

Presenters: Shafner L1, McCue M2, Rubin A2, Dong X2, Hanson E2, Mahableshwarkar A2, Hanina A1, Macek T2

Affiliations: 1AiCure, New York, New York; 2Takeda Development Center Americas, Inc., Deerfield, Illinois

Objective: The need to minimize medication nonadherence is particularly important in central nervous system (CNS) clinical trials. An artificial intelligence (AI) platform was assessed in measuring and increasing medication adherence in subjects with schizophrenia in a Phase 2, randomized, double-blind study.

Design: Subjects in the TAK-063 study who were stable after three or more weeks of inpatient treatment and were discharged were followed up as outpatients for the remainder of the six-week period. Subjects were given devices with the AI application downloaded and asked to use the application for dosing administrations. The primary adherence measure for the study was based on scheduled pill counts; AI platform adherence data were tested for exploratory purposes.

Results: The AI platform was used by 26 subjects for 372 subject days; 744 adherence parameters were collected. Five subjects discontinued early (19.2%). For subjects who completed the trial, mean (standard deviation [SD]) cumulative adherence rates based on visual confirmation of drug ingestion (AI application) and on pill count were 82.5 percent and 99.7 percent, respectively. The mean time to use the AI platform was 86.8 seconds per pill.

Conclusion: Subjects with acutely exacerbated schizophrenia who were eligible for discharge from the inpatient setting and who completed the study demonstrated high rates of adherence using the mobile AI application. Subjects were able to easily use the technology. Use of the platform did not appear to increase the dropout rate. This study demonstrates the feasibility of using AI platforms to ensure high adherence, provide reliable adherence data, and rapidly detect nonadherence in CNS trials.

Disclosures/funding: Adam Hanina and Laura Shafner are employees of AiCure, New York, New York, and consultants to Takeda. Xinxin Dong, Elizabeth Hanson, Thomas A. Macek, Atul Mahableshwarkar, and Maggie McCue are employees of Takeda Development Center Americas, Inc., Deerfield, Illinois. Anne Rubin is a former employee of Takeda.

 

Optimizing the mobile clinical trial: data collection through a web-based platform in telemedicine protocol

Presenters: Shaw M1, Song G2, Charvet L1, Beringer J3, Waibel U3

Affiliations: 1NYU School of Medicine, New York, New York; 2Stony Brook Medicine, Stony Brook, New York; 3Berisoft Corp., Redwood City, California

Objective: A web-based platform was employed to administer cognitive testing and self-reported mood in a clinical trial where the intervention was administered remotely to participants at home through telerehabilitation.

Design: Using a remotely supervised telemedicine protocol, our study tested the benefit of transcranial direct current stimulation (tDCS) for participants with multiple sclerosis (MS). Participants completed daily assessments along with their treatments through a study-provided laptop. To avoid paperwork and to access data in real-time, we utilized the ERTSLab online platform, administrable from any computer with an internet connection. Cognitive functioning was measured by the Attention Networks Test-Interaction (ANT-I) and self-reported mood was measured using the Positive and Negative Affect Schedule (PANAS) taken from the Cognition Library of ERTSLab.

Results: For the ANT-I, excellent form reliability was found between ERTSLab and the PC-installed ePrime version. Form reliability was excellent (r=0.90). Using the ERTSLab version, in our MS sample we found significant improvement in alerting scores for the treatment (n=6) versus control (n=4) conditions (p=0.031). For the PANAS, the self-reported mood measure, online versus standard forms were similarly validated (r=0.90). For the study, 13 participants completed a daily PANAS assessment along with their treatment sessions, and we found a trend toward statistically significant improvement in positive (p=0.001) and negative (p=0.70) affects for those in the treatment versus control condition.

Conclusion: Online platforms such as ERTSLab provide a mobile platform to administer cognitive assessment and self-reported symptom inventories in study designs where participants are assessed in remote locations including the use of telerehabilitation.

Disclosures/funding: None reported.


PATIENT ASSESSMENT AND ADHERENCE

Are there reporting differences at screening on the Montgomery-Äsberg Depression Rating Scale (MADRS) between older and younger adults when using remote assessments?

Presenters: Yavorsky C, Seglund P, Engelhardt N, McNamara C

Affiliations: All authors employees of Cronos CCS, Lambertville, New Jersey

Objective: Our objective was to assess whether differences exist between two age groups of patients (older adults and younger adults) at screening by comparing mean MADRS scores and individual item differences.

Design: An independent samples t-test was performed to compare means at screening for the two age groups (younger adults and older adults). Analysis was conducted using SPSS 21.0 for Windows. Specific item analysis was also conducted to determine if there were effects related to symptom domain.

Results: There was a significant difference (t[55]=51.45, p<0.0001) between the two groups, with the younger patient cohort having significantly higher scores at screening than the older patient cohort.

Conclusion: There were significant differences between older and younger patient cohorts at screening, with older adults reporting less severity than younger adults. There is a paucity of research around phone versus face-to-face interviewing in older adults in terms of rapport and acceptability. These findings suggests there may be real differences that merit further study to determine if the effect is related to characteristics of the geriatric depression population or the methods and expectations about telephone interviewing in this patient group.

Disclosures/funding: All authors are employees of Cronos CCS, Lambertville, New Jersey, and report no conflicts of interest.

 

Cultural adaptation of the Virtual Reality Functional Capacity Assessment Tool (VRFCAT) for use in the UK and Canada

Presenters: Atkins A1, Saxby B1,2 Kelly S1, Hamby M1, Roux P1, Gonzalez M1, Madden J1, Stankovic M1, Keefe R1,3

Affiliations: 1NeuroCog Trials, Durham, North Carolina; 2 Institute for Ageing, Newcastle University, Newcastle, United Kingdom; 2 Duke University Medical Center, Durham, North Carolina

Objective: The Virtual Reality Functional Capacity Assessment Tool (VRFCAT) uses a computer-simulated environment to assess a subject’s ability to complete instrumental activities associated with a shopping trip. With six alternate forms to prevent practice effects, the VRFCAT has demonstrated high test-retest reliability and has shown sensitivity to functional impairment in schizophrenia. Originally developed in English for the United States, we describe cultural adaptation of the VRFCAT for use in the United Kingdom and Canada.

Design: Our method for cultural adaptation followed recommendations of the International Society for Pharmacoeconomics and Outcomes Research (ISPOR). Cultural adaptation was completed independently by two cultural experts in each region using concept sheets, screenshots, and additional task materials. Discrepancies were reconciled through discussion and ultimate consensus among reviewers and content experts.

Results: Culturally adapted test versions were created based on thorough review of feedback received from in-country reviewers. Culturally specific graphic and audio content were created, including design of virtual environments, objects and icons, and professional recording of voiceovers. Other culturally specific changes included revisions of wording (e.g., changing “apartment” to “flat” for the United Kingdom, currency, and prices. Customization of food items and recipes were required to account for cultural differences in item frequency and familiarity. More extensive adaptation was required for the United Kingdom, including the street scene and bus appearance.

Conclusion: Cross-cultural adaptation of the VRFCAT revealed significant variations across English-speaking cultures and highlighted the importance of appropriate adaptation of functional assessments for use in multinational trials.

Disclosures/funding: AS Atkins, BK Saxby, SE Kelley, M Hamby, P Roux, M Gonzalez, J Madden, and M Stankovic are full-time employees of NeuroCog Trials. RSE Keefe is owner and CEO of NeuroCog Trials, the company that developed the VRFCAT as a proprietary instrument and provides commercial distribution and support services. Support provided by National Institute of Mental Health under award R44MH084240.

 

Development of neurophysiological-based biomarkers for neurodegenerative disease and psychiatric disorders using EEG

Presenters: Waninger S1, Berka C1, Stikic M1, Korszen S1, Salat D2, Verma A3

Affiliations: 1Advanced Brain Monitoring, Inc., Carlsbad, California; 2MGH/MIT/HMS Athinoula A. Martinos Center for Biomedical Imaging, Department of Radiology, Massachusetts General Hospital, Charlestown, Massachusetts ; 3Biogen, Cambridge, Massachusetts

Background: Successful drug development for neurodegenerative diseases and psychiatric disorders requires objective, reliable, and accurate measures to evaluate disease progression and therapeutic efficacy. The potential of neurophysiological biomarkers using electroencephalography (EEG) has been highlighted in ongoing studies at Advanced Brain Monitoring on Parkinson’s disease (PD), Alzheimer’s disease (AD), mild cognitive impairment (MCI), mood disorders, and posttraumatic stress disorder (PTSD).

Design: Using our wireless B-Alert X24 system with a standard 10–20 montage, we acquired and analyzed EEG data both during resting state and neurocognitive tasks designed to activate the neural circuits involved in attention, memory, and emotion and elicit event-related potentials (ERPs). For resting state, data were acquired for five minutes with eyes closed (EC) and five minutes with eyes open (EO). These data were decontaminated and converted from the time domain to the frequency domain using Fast Fourier Transform (FFT) to calculate power spectral densities (PSD) grouped into the standard EEG bandwidths (delta, theta, alpha, beta, and gamma). Comparison of PSD from AD and MCI patients to healthy, age-matched controls indicated distinguishing features, particularly the “slowing” of EEG exemplified by an increase in slow wave bands and a decrease in fast power that is typically observed in patients with cognitive decline. The source of the abnormal EEG in MCI patients was localized to the middle and superior temporal gyrus and fusiform gyrus using low resolution electromagnetic tomography (LORETA). Variables extracted from the resting state data (i.e., absolute PSDs, relative PSDs, and wavelets) were grouped together into a feature vector, and the most discriminative variables were selected to construct linear discriminate function analysis (lDFA) models.

Results: Application of lDFA to the MCI dataset resulted in high accuracy and specificity using auto-validation or leave-one-out cross-validation. Although a great deal of information was derived from resting state data, we captured engagement of neural circuits during neurocognitive tasks that stimulate and elicit neural patterns associated with attention, memory, and emotion. One such task, three-choice vigilance (3CVT), is a visual choice reaction time task designed to measure sustained attention and target detection. EEG data acquired during the 3-CVT task indicated significantly longer peak latency in the parietal N2 component of the ERP in a PTSD cohort compared to healthy control cohort.

Conclusion: These data provided further evidence that PTSD patients have increased difficulty with attentional resources, and identified a potential biomarker for PTSD disease progression. Both resting state and event-related EEG data have potential for use as cost-efficient, noninvasive, pharmacodynamic endpoints in neurodegenerative disease and psychiatric disorder clinical trials of experimental therapeutics.

Disclosures/funding: None reported.

 

Discrepancies between CGI-S score and PANSS item level scores—an exploratory analysis

Presenters: Kott A, Daniel D,

Affiliations: Bracket Global, Wayne, Pennsylvania

Objectives: The objective of this study was to identify and characterize factors associated with discrepancies in the rating of the Clinical Global Impression-Severity (CGI-S) scale versus the Positive and Negative Syndrome Scale (PANSS) in a large database of schizophrenia clinical trials rating data.

Design: We retrospectively analyzed 67,698 subject visits from 15 multicenter, double-blind, placebo-controlled schizophrenia trials. CGI-S versus PANSS item level discrepancies were operationally defined as ratings where CGI-S score was at least three points below the highest individual PANSS item score. Using univariate logistic regression models, we assessed the effect of overall and individual symptom severity, PANSS total versus CGI-S discrepancy, region, study type, visit type, and changes in raters on the presence of the discrepancies.

Results: In this database, the prevalence of CGI-S versus PANSS item level discrepancies was 1.33 percent (904/67,698). Visit type, study type, and region had significant effects on the presence of item discrepancies. Increased item severity and different raters significantly increased the odds of PANSS item CGI-S discrepancies, while the odds significantly decreased with increasing overall severity.

Conclusion: We found a prevalence of more than one percent of PANSS item level versus CGI-S discrepancies. In their majority, these discrepancies were distinctly different from PANSS total score versus CGI-S discrepancies. We identified multiple factors significantly impacting on the presence of these item level discrepancies. We conclude that algorithms assessing possible PANSS-CGI discrepancies should consider individual item severity in addition to total scores. Further research is necessary to replicate and understand better the findings.

Disclosures/funding: Both authors are full time employees of Bracket.

 

An electronic (eCOA)-prompted Yale Global Tic Severity Scale (YGTSS) with blinded internal scoring

Presenters: Busner J1, Farber R2, O’Brien C2, Liang G2, Scahill L3, Coffey B4

Affiliations: 1Penn State College of Medicine, Department of Psychiatry, Hershey, Pennsylvania, and Bracket Global, Wayne, Pennsylvania; 2Neurocrine Biosciences, Inc., San Diego, California; 3Emory University, Atlanta, Georgia; 4Icahn School of Medicine at Mount Sinai School, New York, New York

Objective: The YGTSS is a gold-standard tic severity assessment in Tourette’s syndrome (TS) studies. To improve ratings quality, we developed an eCOA-prompted YGTSS that provided rating guidance, captured rater scores, and generated rater-blinded algorithm-derived scores as quality checks.

Design: An eCOA-prompted YGTSS was developed with TS experts. The eCOA YGTSS ensured correct navigation through the scale and assisted raters by displaying earlier endorsed tics when needed for motor and phonic severity ratings. The scale included algorithms for a second set of rater-blinded “tandem” scores. The scale is being piloted in two ongoing, placebo-controlled, multisite TS trials, one pediatric and one adult, with rater scores serving as efficacy data.

Results: There were 99 visits by 20 raters that were completed at the time of the analysis.  Correlations between rater and computer scores were high for each of the 10 YGTSS severity scores (range: 0.74–0.91, all p values <0.0001); for the Total Tic Score (TTS) (primary efficacy measure), the correlation was 0.95 (p<0.0001).  The mean rater versus computer TTS scores were almost identical (28.8 and 28.5, respectively, NS). The findings did not differ between pediatric and adult subjects.

Conclusion: Our internal scoring algorithms correlated significantly with all rater-selected motor, phonic, and TTS scores, with the latter nearly identical. The work provides preliminary validation of our algorithms and supports the feasibility of the approach. In a risk-based monitoring model, less trained raters whose scores deviate significantly from those of the internal algorithm might be selected for additional scrutiny and intervention.

Disclosures/funding: Dr. Busner is an employee of Bracket Global, Wayne, Pennsylvania. Drs. Farber, O’Brien, and Liang are employees of Neurocrine Biosciences, Inc., San Diego, California. Dr. Scahill is a consultant to Roche, Bracket, Neuren Pharmaceuticals, MedAdvante, Coronado Biosciences, Inc., and Supernus Pharmaceuticals, Inc.; has received research support from the United States National Institute of Mental Health (NIMH) and Department of Defense; and has received royalties from Oxford and Guilford. Dr. Coffey has received honoraria from American Academy of Child and Adolescent Psychiatry; has received research support from Neurocrine, NIMH, Shire, and Catalyst Pharmaceuticals Inc.; serves on the advisory board, received Center of Excellence Funding, and is on the speakers bureau of Tourette Association of America; serves on the advisory board and receives research support from Auspex Pharmaceuticals; and serves on the advisory board of Genco Sciences.

 

The electronic self-report of the C-SSRS (eC-SSRS) places little burden on patients initially and even less with repeated administrations

Presenters: Yamamoto R1, Durand E1, Lima V1, Christopher D1, Yershova K2, Dallabrida S1

Affiliations: 1ERT Corp., New York, Neew York2; Columbia University/New York State Psychiatric Institute, New York, New York

Objective: A key factor to consider for instruments that measure suicidal ideation and behavior (SIB) is the duration of patient completion time. The Columbia Suicide Severity Rating Scale (C-SSRS) (interviewer completed) and electronic C-SSRS (eC-SSRS) (self-report) are sourced by the United States Food and Drug Administration (FDA) in the SIB Industry Guidance as recommended assessments. This study examined patient completion time for the tablet-based eC-SSRS over time.

Design: Over 2,000 subjects with substance abuse disorder completed the tablet-based eC-SSRS assessment at each clinic visit in a trial. Data from the eC-SSRS since last contact version were analyzed.

Results: Completion times for patients who scored negative for SIB using the tablet version of the e-C-SSRS were 1.15±1.33min at Visit 1, 0.84±0.89min at Visit 7, and 0.73±0.79min at Visit 12. There was a significant decrease in the time it took patients to complete the e-CSSRS across Visits 1 through 12 General Linear Model (mixed) (F=5809.25, p<0.0001). ANOVA repeated measures on all 12 visits (F=7.7, p<0.0001) showed that the time to complete the eC-SSRS decreased across visits. A t-test between Visit 1 and Visit 12 (p<0.0001, t=4.8) indicated that it took significantly less time for subjects to complete the eC-SSRS on Visit 12 versus Visit 1.

Conclusion: These data showed that there was little burden on subjects to complete the tablet eC-SSRS, and this burden decreased with repeated administrations. This is a particularly important consideration in lengthy studies with many repeat collections of an SIB assessment.

Disclosures/funding: Drs. Yamamoto, Durand, and Dallabrida and Ms. Lima and Mr. Christopher are employees of ERT Corp., New York, New York. Dr. Yershova is an employee of Columbia University/New York State Psychiatric Institute, New York, New York.

 

Evaluating the use of an artificial intelligence platform on mobile devices to measure adherence in subjects with an acute exacerbation of schizophrenia

Presenters: Shafner L1, McCue M2, Rubin A2, Dong X2, Hanson E2, Mahableshwarkar A2, Hanina A1, Macek T2

Affiliations: 1AiCure, New York, New York; 2Takeda Development Center Americas, Inc., Deerfield, Illinois

Objective: The need to minimize medication nonadherence is particularly important in central nervous system (CNS) clinical trials. An artificial intelligence (AI) platform was assessed in measuring and increasing medication adherence in subjects with schizophrenia in a Phase 2, randomized, double-blind study.

Design: Subjects in the TAK-063 study who were stable after three or more weeks of inpatient treatment and were discharged were followed up as outpatients for the remainder of the six-week period. Subjects were given devices with the AI application downloaded and asked to use the application for dosing administrations. The primary adherence measure for the study was based on scheduled pill counts; AI platform adherence data were tested for exploratory purposes.

Results: The AI platform was used by 26 subjects for 372 subject days; 744 adherence parameters were collected. Five subjects discontinued early (19.2%). For subjects who completed the trial, mean (standard deviation [SD]) cumulative adherence rates based on visual confirmation of drug ingestion (AI application) and on pill count were 82.5 percent and 99.7 percent, respectively. The mean time to use the AI platform was 86.8 seconds per pill.

Conclusion: Subjects with acutely exacerbated schizophrenia who were eligible for discharge from the inpatient setting and who completed the study demonstrated high rates of adherence using the mobile AI application. Subjects were able to easily use the technology. Use of the platform did not appear to increase the dropout rate. This study demonstrates the feasibility of using AI platforms to ensure high adherence, provide reliable adherence data, and rapidly detect nonadherence in CNS trials.

Disclosures/funding: Adam Hanina and Laura Shafner are employees of AiCure, New York, New York, and consultants to Takeda. Xinxin Dong, Elizabeth Hanson, Thomas A. Macek, Atul Mahableshwarkar, and Maggie McCue are employees of Takeda Development Center Americas, Inc., Deerfield, Illinois. Anne Rubin is a former employee of Takeda.

 

NetraMark: predicting placebo and drug response for pharma

Presenters: Geraci J

Affiliations: NetraMark Corp., Ontario, Canada

Objective: We wished to showcase precision medicine through novel machine learning models based on molecular and psychiatric scale data that can impact clinical trial success through placebo response and drug efficacy prediction.

Design: Utilizing data sets, we discovered models of placebo response robust enough to reproduce (within the specific treatment paradigm of a trial) and models that demonstrated the patient population purification process for predicting response in a precise way.

Results: We found two models of placebo response that had an accuracy of over 85 percent. For some situations we were able to predict if someone was a placebo responder with an accuracy of over 95 percent. Clinical scales alone were used for one of these models. We also based models on miRNA expression data that were capable of accurately predicting response for a particular subpopulation of major depression patients. This provided evidence that our technology is capable of dealing with heterogeneous patient populations. This could be valuable during Phase 3 scenarios for drug protocols that only work for a portion of the treatment population.

Conclusion: NetraMark developed a machine learning system to help pharmaceutical companies who must deal with complex patient populations. NetraMark accomplished this by providing accurate models of placebo and drug response. In situations where the efficacy of a drug is limited to a subpopulation whose effect could be “washed out” by the United States Food and Drug Administration (FDA) statistical paradigm, we provided what we call a Patient Population Purification process. This process is capable of predicting who these patients are with a high level of accuracy. Our unique system has the ability to help companies get potential medications on the market through our approach to precision medicine.

Disclosures/funding: We are now working with several pharmaceutical companies. At this time, we are unable to disclose all our current relationships.

 

The NetSCID: a validated web-based adaptive version of the SCID that improves diagnostic accuracy in mental health research

Presenters: Brodey BB1, Zweede L1

Affiliations: 1TeleSage, Inc., Chapel Hill, North Carolina

Background: The Structured Clinical Interview for the Diagnostic and Statistical Manual of Mental Disorders (SCID, DSM) is the gold standard for research-based mental health diagnoses. It is the most widely used comprehensive tool for assessing DSM diagnoses. Its direct adherence to DSM criteria provides strong test-retest and high interrater reliability; however, administration of the full research version averages two hours and requires considerable clinician training, making it impractical for many protocols.

Objective: Our objective was to develop and validate a highly configurable and secure web-based version of the SCID—the NetSCID— thereby making the gold standard in mental health diagnostics available for clinical trials.

Design: The validation research included 24 clinicians who administered the SCID to 230 participants that completed the paper SCID and/or the NetSCID. Data-entry errors, branching errors, and clinician satisfaction were quantified.

Results: Ninety-seven percent of error rates occurred among clinicians administering the paper SCID; all errors that occurred when utilizing the NetSCID were subsequently corrected by our programmers. Administration time was reduced by over 30 percent. Clinicians found it easier to administer (p< 0.05), easier to navigate (p< 0.05), and simpler to score (p< 0.01). Ninety percent of clinicians preferred the NetSCID to its paper counterpart.

Conclusion: Because of its unique ability to record symptoms in a standardized database, the NetSCID facilitates characterization of patient populations and assists with identification of sub-groups that may respond to interventions. Adoption of this tool has shown to improve diagnostic accuracy and will increase the power in central nervous system (CNS) clinical trials worldwide.

Disclosures/funding: This research was supported in part by a grant from the National Institutes of Mental Health, which was awarded to TeleSage, Inc. Dr. Brodey is Chief Executive Officer of TeleSage, Inc., and Lisa Zweede is employed by TeleSage, Inc. as a clinical research specialist.

 

Patients with psychiatric diagnoses indicate willingness to report suicidal ideation and behavior more honestly by self-report than in face-to-face interviews

Presenters: Yamamoto R1, Durand E1, Khurana L1, Tuller J2, Yershova K2, Dallabrida S1

Affiliations: 1ERT Corp., New York, New York; 2Columbia University/New York State Psychiatric Institute, New York, New York

Objective: Detection of honest/accurate patient experiences with suicidal ideation and behavior (SIB) is critical to patient safety and to determining drug efficacy and safety. Herein, psychiatric patients were asked about the situations in which they would be most likely to honestly report their SIB.

Design: Patients (recruited via Clinical Connection) who reported at least one psychiatric diagnosis (depression, generalized anxiety disorder, bipolar disorder, alcohol/drug addiction, or schizophrenia) and previously answered questions at a clinic visit about having thoughts of suicide/acts of self-harm were asked in which situation they would be most likely to honestly report their SIB.

Results: Ninety-two percent of the patients (n=63, ages 18–60, 68% female) indicated that they would be more likely to honestly report SIB using electronic self-report assessments versus 74 percent of these same patients who indicated they would be likely to be honest during an in-person interview at a clinic visit. In psychiatric patients who truly exhibit SIB, the relative risk (ratio of proportion of risk between groups) that clinicians will miss SIB during face-to-face interviews is 3.37 times higher than using electronic self-report assessments (p=0.0225, 95% confidence interval 1.1866–9.5730). The odds ratio (probability of missing SIB) that a patient would not be honest reporting SIB in-person is 4.2 times greater than the odds that a patient would not be honest reporting SIB electronically (p<0.02).

Conclusion: These data suggest that psychiatric patients who have experience completing SIB assessments are more likely to respond honestly to future questions about SIB via electronic self-report assessments.

Disclosures/funding: Drs. Yamamoto, Durand, and Dallabrida and Ms. Khurana and Tuller are employees of ERT Corp. Dr. Yershova is an employee of Columbia University/New York State Psychiatric Institute, New York, New York.

 

Translation of the National Institutes of Health Stroke Scale (NIHSS) list of words: methodology and challenges

Presenters: Vasarri S1, Veal L1, Anfray C2, Emery M2

Affiliations: 1Mapi Language Services, Lyon, France; 2Mapi Research Trust, Lyon, France

Objective: The NIHSS was developed to assess stroke severity across 11 categories. Dysarthria is evaluated with a list of six terms, each one exploring different lips and tongue movements: mama, tip-top, fifty-fifty, thanks, huckleberry, and baseball player. The objective of our study was to present the methodology used to translate these words into Canadian French, Bulgarian, Korean, and American Spanish, and the resulting outcomes. The aim was not to find conceptual equivalents but words testing the same difficulties in articulation and using similar phonological characteristics as the original terms.

Design: In each country, a thorough translation was performed with a neurologist and a speech therapist. The suggested words were then validated by an expert panel.

Results: Only one word—mama—was translated literally in all languages, the pronunciation being similar. The translation of tip-top used satisfactory phonological equivalents in all languages. Thanks was difficult to translate because the sound (“th” [ð]) does not exist in the target languages. Equivalent words allowing measuring the weakness of the tongue were used. Fifty-fifty was rendered completely differently in Korean, since the fricative sound (f) does not exist in this language. The Korean translation of huckleberry could not contain all the original sounds. As for baseball player, this terms was referred to as baseball in Canadian French and Bulgarian.

Conclusion: Finding phonological equivalents to terms used to assess dysarthria was challenging and needed the collaboration of speech therapists and neurologists in each target language. Korean was the most challenging lanaguage due mainly to the absence of equivalent sounds.

Disclosures/funding: None reported.


PLACEBO RESPONSE

Factors associated with response rate in inpatient schizophrenic participants

Presenters: Tireman E, Templeton K, Kakar R

Affiliations: All authors are from Segal Institute for Clinical Research, Fort Lauderdale, Florida.

Background: Identifying appropriate patients for inpatient clinical trials can be challenging due to a number of factors. Placebo response and early response are factors that hinder the outcomes of clinical trials and often lead to many failed trials or lack of separation data. Demographics, number of prior hospitalizations, and previous participation in clinical trials may all influence how patients respond to treatment and their success in completing a trial. Currently, there limited research assessing factors associated with early response patterns and placebo response in relation to level of change on primary efficacy outcome measures. Recent research has been evaluating the relationship between demographics (gender and body mass index [BMI]), lifetime hospitalizations, and placebo response. The ideal patient has been described as younger in age with limited number of lifetime hospitalizations or previous clinical trial experience. Current discussions among rating professionals suggest that older male subjects with a higher BMI, greater number of lifetime hospitalizations and/or exposure to clinical trial procedures will demonstrate a greater difference in Positive and Negative Syndrome Scale (PANSS) scores beginning earlier in the trial. These individuals are likelier to have a higher placebo response rate and often times do not complete trials. Identifying those patients who show a tendency to respond early to treatment may assist research sites in recruiting subjects of a different population as well as improving overall success of clinical trials.

Design: Factors including subjects’ age, sex, BMI, prior hospitalizations, and PANSS scores were collected on 90 patients with schizophrenia who were experiencing an acute exacerbation of symptoms and presented for inclusion in an eight-week inpatient trial. These factors were analyzed to determine whether there is a correlation in response rate in PANSS scores in relation to these factors.

Results: Correlational analyses were used to examine the relationship between demographic factors (age, sex, BMI, and prior hospitalizations) and response rate in PANSS scores. There was no significant relationship found between demographic factors and percent change in total PANSS score. However, preliminary correlational analyses indicated an inverse relationship between PANSS early onset schizophrenia (EOS) scores and overall percent change (r= -0.618, p=0.000). This suggests subjects with higher EOS scores showed less change overall. Additionally, an inverse relationship between PANSS ET score and overall percent change (r= -.397, p=0.009) also suggesting a higher ET score indicated a lower overall percent change.

Conclusion: These data indicated the demographic factors of our subset of patients did not significantly relate to their overall response rate in the trials. Information that would assist in a stronger analysis of these factors includes a greater number of subjects for analysis, total number of previous clinical trial exposure, and site outcome data. This topic of discussion is particularly relevant to the continued success in recruitment and retention of subjects and overall positive trial outcome data.

Financial disclosure/funding: None reported.

 

NetraMark: predicting placebo and drug response for pharma

Presenters: Geraci J

Affiliations: NetraMark Corp., Ontario, Canada

Objective: We wished to showcase precision medicine through novel machine learning models based on molecular and psychiatric scale data that can impact clinical trial success through placebo response and drug efficacy prediction.

Design: Utilizing data sets, we discovered models of placebo response robust enough to reproduce (within the specific treatment paradigm of a trial) and models that demonstrated the patient population purification process for predicting response in a precise way.

Results: We found two models of placebo response that had an accuracy of over 85 percent. For some situations we were able to predict if someone was a placebo responder with an accuracy of over 95 percent. Clinical scales alone were used for one of these models. We also based models on miRNA expression data that were capable of accurately predicting response for a particular subpopulation of major depression patients. This provided evidence that our technology is capable of dealing with heterogeneous patient populations. This could be valuable during Phase 3 scenarios for drug protocols that only work for a portion of the treatment population.

Conclusion: NetraMark developed a machine learning system to help pharmaceutical companies who must deal with complex patient populations. NetraMark accomplished this by providing accurate models of placebo and drug response. In situations where the efficacy of a drug is limited to a subpopulation whose effect could be “washed out” by the United States Food and Drug Administration (FDA) statistical paradigm, we provided what we call a Patient Population Purification process. This process is capable of predicting who these patients are with a high level of accuracy. Our unique system has the ability to help companies get potential medications on the market through our approach to Precision Medicine.

Disclosures/funding: We are now working with several pharmaceutical companies. At this time, we are unable to disclose all our current relationships.


RATER TRAINING AND ASSESSMENT

Understanding rater preferences in services used to increase the reliability of clinical trials: a multi-national survey

Presenters: Komorowsky A1, DiCindio L1, Rock C1, Lobb J1, Avrumson R1, Carbo M1, Baldwin K1, Rohleder L1, Lytle D1, Murphy M1

Affiliations: 1Worldwide Clinical Trials, King of Prussia, Pennsylvania

Objective: Industry literature supports the use of independent, expert clinicians to train and monitor data quality for psychiatric/neurological assessments in order to reduce bias and variability in outcomes data. As the number of clinical trials continues to increase, so does the frequency with which site raters undergo training and monitoring using different modalities and standards. The authors investigated raters’ perceptions and preferences with common rater training and data surveillance programs as well as the potential impact of rater engagement on data quality.

Design: A nonprobability-based, purposive survey was employed in which site staff engaged in interventional psychiatry/neurology studies was contacted from the Worldwide Clinical Trials database of site raters and coordinators. Over 2,000 surveys were deployed using a proprietary survey engine.

Results: Reponses from the survey revealed key rater preferences in methods of training and surveillance. Specific preferences were found for training via video demonstration with a practice quiz as well as preferences for in-person investigators’ meetings. Data surveillance preferences were found for some methodologies, including source document review. The survey also identified pre-conceived perceptions of the raters regarding the purpose of rater training and surveillance.

Conclusion:  Understanding site rater preferences is paramount to enriching the current rater training and data monitoring methods, which can directly impact study outcomes. Implications for training and quality assurance methodology were also outlined.

Disclosures/funding: All authors are full-time employees of Worldwide Clinical Trials, King of Prussia, Pennsylvania, and have no conflicts of interest.


TRIAL METHODOLOGY/STUDY PROTOCOL

Could a paradigm shift in consumptive disorders research benefit patients and sponsors alike? A smoking cessation case study reinforces “the manner in which we look at things determines what we see!”

Presenters: Wilcox C1, Oskooilar N2, Morrissey J3, Rosenberg D2, Tong M1 Henry M2, Badgett L3, Grosz D3

Affiliations: 1Pharmacology Research Institute [PRI] Newport Beach, California; 2PRI Los Alamitos, California; 2PRI Encino, California

Objective: Our objective was to investigate the potential benefit of re-analyzing data (i.e., “changing the paradigm”) currently used for “Go vs. No-go” clinical development decisions.

Rationale: The evaluation of potential new medicines for nearly all central nervous system (CNS) indications focuses on symptom severity reduction (usually ?50%), not total elimination. The current regulatory threshold for successful efficacy in smoking cessation is a total (100%) and sustained elimination. This is an arbitrarily high bar. Additionally, we question how many efficacious smoking cessation treatments may have fallen victim to Type II errors using these criteria.

Design: We conducted a randomized, double-blind, 12-week, placebo-controlled, dose selection study. Seventy-five smokers were screened and 59 were randomly assigned to one of three treatment groups: lorcaserin 10mg once daily, lorcaserin 10mg twice daily, or placebo in a 3:3:2 ratio. The primary efficacy endpoint was the end-expiratory carbon monoxide (CO)-confirmed continuous abstinence rate (CAR) from Study Weeks 9 to 12, defined as zero reported smoking via Nicotine Use Inventory (NUI) with exhaled CO measurement 10ppm or less.

Results: None of our initial results demonstrated any statistically significant findings. Analyses based on revised success criteria produced highly significant results. Using either NUI=0 or CO?10ppm, high-dose lorcaserin demonstrated significantly better efficacy when compared to placebo (p<0.02) and low dose lorcaserin (p<0.005). When we looked at harm/risk reduction, defined by NUI values of 5 or less and a CO value of 10ppm or less, hereto lorcaserin 20mg was significantly more efficacious than placebo (p<0.02) and low-dose lorcaserin (p<0.01).

Conclusion: “Don’t let the perfect be the enemy of the good” (Voltaire, 1694–1778). Although lorcaserin 10mg once daily and twice daily did not demonstrate statistically significant results via pre-specified protocol criteria of 100-percent sustained cessation, a different conceptual and statistical lens produced quite robust results. When re-analyzed utilizing our harm reduction criteria, high-dose lorcaserin was statistically significantly superior to both placebo and low-dose lorcaserin at multiple time points, including Endpoint/Week 12. High-dose lorcaserin appears to have clinical and commercial potential as an efficacious smoking cessation agent. We believe it warrants further investigation, with or without a paradigm shift in the definition of successful treatment.

Disclosures/funding: The authors have no conflicts of interest relevant to the study. Funding for this retrospective analysis was provided internally by Pharmacology Research Institute.

 

Discrepancies between CGI-S score and PANSS item level scores—an exploratory analysis

Presenters: Kott A, Daniel D,

Affiliations: Bracket Global, Wayne, Pennsylvania

Objectives: The objective of this study was to identify and characterize factors associated with discrepancies in the rating of the Clinical Global Impression-Severity (CGI-S) scale versus the Positive and Negative Syndrome Scale (PANSS) in a large database of schizophrenia clinical trials rating data.

Design: We retrospectively analyzed 67,698 subject visits from 15 multicenter, double-blind, placebo-controlled schizophrenia trials. CGI-S versus PANSS item level discrepancies were operationally defined as ratings where CGI-S score was at least three points below the highest individual PANSS item score. Using univariate logistic regression models, we assessed the effect of overall and individual symptom severity, PANSS total versus CGI-S discrepancy, region, study type, visit type, and changes in raters on the presence of the discrepancies.

Results: In this database, the prevalence of CGI-S versus PANSS item level discrepancies was 1.33 percent (904/67,698). Visit type, study type, and region had significant effects on the presence of item discrepancies. Increased item severity and different raters significantly increased the odds of PANSS item CGI-S discrepancies, while the odds significantly decreased with increasing overall severity.

Conclusion: We found a prevalence of more than one percent of PANSS item level versus CGI-S discrepancies. In their majority, these discrepancies were distinctly different from PANSS total score versus CGI-S discrepancies. We identified multiple factors significantly impacting on the presence of these item level discrepancies. We conclude that algorithms assessing possible PANSS-CGI discrepancies should consider individual item severity in addition to total scores. Further research is necessary to replicate and understand better the findings.

Disclosures/funding: Both authors are full time employees of Bracket.

 

Evaluating the use of an artificial intelligence platform on mobile devices to measure adherence in subjects with an acute exacerbation of schizophrenia

Presenters: Shafner L1, McCue M2, Rubin A2, Dong X2, Hanson E2, Mahableshwarkar A2, Hanina A1, Macek T2

Affiliations: 1AiCure, New York, New York; 2Takeda Development Center Americas, Inc., Deerfield, Illinois

Objective: The need to minimize medication nonadherence is particularly important in central nervous system (CNS) clinical trials. An artificial intelligence (AI) platform was assessed in measuring and increasing medication adherence in subjects with schizophrenia in a Phase 2, randomized, double-blind study.

Design: Subjects in the TAK-063 study who were stable after three or more weeks of inpatient treatment and were discharged were followed up as outpatients for the remainder of the six-week period. Subjects were given devices with the AI application downloaded and asked to use the application for dosing administrations. The primary adherence measure for the study was based on scheduled pill counts; AI platform adherence data were tested for exploratory purposes.

Results: The AI platform was used by 26 subjects for 372 subject days; 744 adherence parameters were collected. Five subjects discontinued early (19.2%). For subjects who completed the trial, mean (standard deviation [SD]) cumulative adherence rates based on visual confirmation of drug ingestion (AI application) and on pill count were 82.5 percent and 99.7 percent, respectively. The mean time to use the AI platform was 86.8 seconds per pill.

Conclusion: Subjects with acutely exacerbated schizophrenia who were eligible for discharge from the inpatient setting and who completed the study demonstrated high rates of adherence using the mobile AI application. Subjects were able to easily use the technology. Use of the platform did not appear to increase the dropout rate. This study demonstrates the feasibility of using AI platforms to ensure high adherence, provide reliable adherence data, and rapidly detect nonadherence in CNS trials.

Disclosures/funding: Adam Hanina and Laura Shafner are employees of AiCure, New York, New York, and consultants to Takeda. Xinxin Dong, Elizabeth Hanson, Thomas A. Macek, Atul Mahableshwarkar, and Maggie McCue are employees of Takeda Development Center Americas, Inc., Deerfield, Illinois. Anne Rubin is a former employee of Takeda.

 

Methodology and management of clinical trials with adaptive designs

Presenters: Laage T, Barag J

Affiliations: Premier Research Group

Objective: Our objective was to present practical techniques for implementing and managing adaptive design clinical trials to preserve study integrity and validity

Design: We reviewed two clinical trial design principles (and their practical application): controlling the chance of erroneous conclusions (minimizing Type I and Type II error by controlling “multiplicity” of data examination at interim assessments and by adjusting for study adaptations) and minimizing operational bias (some or all participants in the study have access to study results by treatment group, influencing the ongoing operations of the study).

Results: Trials are planned using experienced statisticians for analytical methods and computer simulations to evaluate the study design operating characteristics and ensure control of Type I error rate and adequate power. Trials are conducted to minimize operational bias, including the following: masking investigators from knowledge of interim analyses and of study adaptations (sample size increase, longer enrollment, and changed endpoints or randomization ratios); ensuring an adequate firewall (i.e., a process or procedure to ensure and document sufficient restrictions on information flow to control statistical and operational bias); and separating the precise details of the adaptation algorithm from the investigator-shared protocol and placing them in a detailed Statistical Analysis Plan (for institutional review boards [IRBs] and United States Food and Drug Administration [FDA] only).

Conclusion: With careful planning and execution, adaptive trials can realize efficiencies in time and resources.

Disclosures/funding: None reported.

 

NetraMark: predicting placebo and drug response for pharma

Presenters: Geraci J

Affiliations: NetraMark Corp., Ontario, Canada

Objective: We wished to showcase precision medicine through novel machine learning models based on molecular and psychiatric scale data that can impact clinical trial success through placebo response and drug efficacy prediction.

Design: Utilizing data sets, we discovered models of placebo response robust enough to reproduce (within the specific treatment paradigm of a trial) and models that demonstrated the patient population purification process for predicting response in a precise way.

Results: We found two models of placebo response that had an accuracy of over 85 percent. For some situations we were able to predict if someone was a placebo responder with an accuracy of over 95 percent. Clinical scales alone were used for one of these models. We also based models on miRNA expression data that were capable of accurately predicting response for a particular subpopulation of major depression patients. This provided evidence that our technology is capable of dealing with heterogeneous patient populations. This could be valuable during Phase 3 scenarios for drug protocols that only work for a portion of the treatment population.

Conclusion: NetraMark developed a machine learning system to help pharmaceutical companies who must deal with complex patient populations. NetraMark accomplished this by providing accurate models of placebo and drug response. In situations where the efficacy of a drug is limited to a subpopulation whose effect could be “washed out” by the United States Food and Drug Administration (FDA) statistical paradigm, we provided what we call a Patient Population Purification process. This process is capable of predicting who these patients are with a high level of accuracy. Our unique system has the ability to help companies get potential medications on the market through our approach to Precision Medicine.

Disclosures/funding: We are now working with several pharmaceutical companies. At this time, we are unable to disclose all our current relationships.

 

A reanalysis of the effects of protocol design and subject stipend on completion rates in Phase 1 studies in subjects with stable schizophrenia or schizoaffective disorder: Does money matter?

Presenters: Krefetz DG

Affiliations: PRA Health Sciences, Early Development Services, Marlton, New Jersey

Objective: The objective of this study was to evaluate the impact of both protocol design and subject stipend on individual completion in Phase 1 studies in subjects with stable schizophrenia or schizoaffective disorder.

Design: In a poster presented at the central nervous system (CNS) Summit in 2015, the author examined the effect of nine trial design variables on individual completion in 11 Phase 1 trials in subjects with stable schizophrenia or schizoaffective disorder at two clinical research sites during the period from 2009 to 2014. The analysis showed that shortening the length of the inpatient period, increasing the outpatient period, and shortening the longest period between outpatient visits had a positive impact on completion. A criticism of the previous analysis is that it did not include the impact of the subject stipend. Here, we reanalyzed the dataset with the addition of the following variables: the stipend for an inpatient day, the stipend for an outpatient visit, and total stipend divided by the number of days in the study. Stipends are adjusted for inflation using the Consumer Price Index.

Results: The predictive strength of each trial design and stipend variable on subject study completion was presented.

Conclusion: Conclusions were drawn that inform study design for better subject retention.

Disclosures/funding: None reported.

 

Toward a better understanding of informant contributions to schizophrenia trial quality: data from the encenicline cognitive impairment program

Presenters: Nations K1, Rosenthal A1, Spiridonescu L1, Truskowski L1, Quintanilla A1, Prilliman C1, Wise-Rankovic A1, Gibertini M1, Keefe R2, Walker T2, Brannan S3, Hilt D3

Affiliations: 1INC Research, 2NeuroCog Trials, 3FORUM Pharmaceuticals

Objective: Schizophrenia protocols require eligible subjects to bring a reliable informant to confirm symptomatic history. Definitions of “reliable” typically depend on clinical intuition. The current analysis examined informant characteristics as predictors of subject/trial quality in two cognitive impairment in schizophrenia (CIS) Phase 3 trials evaluating encenicline versus placebo.

Design: EVP-6124-015/016 were randomized, double-blind trials (N=1,520). Neither showed a treatment effect on co-primary endpoints, the Measurement and Treatment Research to Improve Cognition in Schizophrenia (MATRICS) Consensus Cognitive Battery (MCCB), and the Schizophrenia Cognition Rating Scale (SCoRS), a measure that includes informant report. A “consistent informant” was required for trial entry and was defined as a person who interacts with the subject at least twice weekly and who could attend site visits. The contract research organization (CRO) medical/clinical team collected data on four informant variables: 1) relationship to subject, 2) duration of relationship, 3) frequency of contact, and 4) weekly time with subject. We examined these factors as predictors of subject quality markers: trial completion, drug adherence, and non-response to placebo on key endpoints.

Results: Frequency of contact between the subjects and their informants significantly predicted trial completion (2=20.01, p<0.001). This quality marker was further supported by an important interaction between treatment group and frequency of contact on the MCCB CFB [F(4,1304)=2.08, p=0.08], showing that subjects in the 2mg group, but not in the placebo or 1mg groups living with their informant were more improved at endpoint than subjects with daily contact.

Conclusion: These data may serve as a source for appropriate and operationally feasible definitions of informant reliability in future CIS protocols. Informant factors are further examined for regional differences and their influence on MCCB/SCoRS relationships.

Disclosures/funding: Funding was provided by FORUM Pharmaceuticals, Waltham, Massachusetts. INC Research, Raleigh, North Carolina, was the CRO responsible for execution of both trials. KN, ASR, LS, LT, AQ, CP, AWR, and MG are employees of INC Research. SB and DH were employees of FORUM at the time of this research. RK is an owner of NeuroCog Trials, Durham, North Carolina, and TW is an employee of NeuroCog Trials.

 

Understanding rater preferences in services used to increase the reliability of clinical trials: a multi-national survey

Presenters: Komorowsky A1, DiCindio L1, Rock R1, Lobb J1, Avrumson R1, Carbo M1, Baldwin K1, Rohleder L1, Lytle D1; Murphy M1

Affiliations: 1Worldwide Clinical Trials, King of Prussia, Pennsylvania

Objective: Industry literature supports the use of independent, expert clinicians to train and monitor data quality for psychiatric/neurological assessments in order to reduce bias and variability in outcomes data. As the number of clinical trials continues to increase, so does the frequency with which site raters undergo training and monitoring using different modalities and standards. The authors investigated raters’ perceptions and preferences with common rater training and data surveillance programs as well as the potential impact of rater engagement on data quality.

Design: A nonprobability-based, purposive survey was employed in which site staff engaged in interventional psychiatry/neurology studies was contacted from the Worldwide Clinical Trials database of site raters and coordinators. Over 2,000 surveys were deployed using a proprietary survey engine.

Results: Reponses from the survey revealed key rater preferences in methods of training and surveillance. Specific preferences were found for training via video demonstration with a practice quiz as well as preferences for in-person investigators’ meetings. Data surveillance preferences were found for some methodologies, including source document review. The survey also identified pre-conceived perceptions of the raters regarding the purpose of rater training and surveillance.

Conclusion:  Understanding site rater preferences is paramount to enriching the current rater training and data monitoring methods, which can directly impact study outcomes. Implications for training and quality assurance methodology were also outlined.

Disclosures/funding: All authors are full-time employees of Worldwide Clinical Trials, King of Prussia, Pennsylvania, and have no conflicts of interest.

Tags: , , , , ,

Category: Assessment Tools, biomarkers, Cognition, Devices, Drug Development, Evaluations, Genetics, Medical Issues, Neurology, Past Articles, Patient Assessment, Proceedings, Psychiatry, Psychopharmacology, Scales, Supplements, Technology, Trial Methodology

Leave a Reply