Designing Robust N-of-1 Studies for Precision Medicine: Simulation Study and Design Recommendations

ABkPowerCalculator: An App to Compute Power for Balanced (AB)k Single Case Experimental Designs

Single case experimental designs are an important research design in behavioral and medical research. Although there are design standards prescribed by the What Works Clearinghouse for single case experimental designs, these standards do not include statistically derived power computations. Recently we derived the equations for computing power for (AB)k designs. However, these computations and the software code in R may not be accessible to applied researchers who are most likely to want to compute power for their studies. Therefore, we have developed an (AB)k power calculator Shiny App (https://abkpowercalculator.shinyapps.io/ABkpowercalculator/) that researchers can use with no software training. These power computations assume that the researcher would be interested in fitting multilevel models with autocorrelations or conduct similar analyses. The purpose of this software contribution is to briefly explain how power is derived for balanced (AB)k designs and to elaborate on how to use the Shiny App. The app works well on not just computers but mobile phones without installing the R program. We believe this can be a valuable tool for practitioners and applied researchers who want to plan their single case studies with sufficient power to detect appropriate effect sizes.

Power analysis for idiographic (within-subject) clinical trials: Implications for treatments of rare conditions and precision medicine

Power analysis informs a priori planning of behavioral and medical research, including for randomized clinical trials that are nomothetic (i.e., studies designed to infer results to the general population based on interindividual variabilities). Far fewer investigations and resources are available for power analysis of clinical trials that follow an idiographic approach, which emphasizes intraindividual variabilities between baseline (control) phase versus one or more treatment phases. We tested the impact on statistical power to detect treatment outcomes of four idiographic trial design factors that are under researchers' control, assuming a multiple baseline design: sample size, number of observations per participant, proportion of observations in the baseline phase, and competing statistical models (i.e., hierarchical modeling versus piecewise regression). We also tested the impact of four factors that are largely outside of researchers' control: population size, proportion of intraindividual variability due to residual error, treatment effect size, and form of outcomes during the treatment phase (phase jump versus gradual change). Monte Carlo simulations using all combinations of the factors were sampled with replacement from finite populations of 200, 1750, and 3500 participants. Analyses characterized the unique relative impact of each factor individually and all two-factor combinations, holding all others constant. Each factor impacted power, with the greatest impact being from larger treatment effect sizes, followed respectively by more observations per participant, larger samples, less residual variance, and the unexpected improvement in power associated with assigning closer to 50% of observations to the baseline phase. This study's techniques and R package better enable a priori rigorous design of idiographic clinical trials for rare diseases, precision medicine, and other small-sample studies.

THE VALUE OF N-OF-1 DATA IN ZOOLOGICAL MEDICINE: A METHODOLOGICAL REVIEW

N-of-1 data are unavoidable in zoological medicine. Accordingly, zoological medicine clinicians and investigators need research techniques that can make use of these data. This article reviews two methodologies for using both observational and experimental N-of-1 data: 1) systematic reviews and meta-analyses of case reports and 2) prospective N-of-1 trials. Systematic reviews of case reports and other observational evidence are formal, unbiased summaries of the clinical characteristics of a particular disease-taxon combination. They offer advantages to narrative reviews by minimizing omission of relevant articles, thereby reducing the potential for mischaracterization of the literature. Meta-analyses are extensions of systematic reviews that quantitatively synthesize the data from the included articles. Although valuable, systematic reviews and meta-analyses of case reports can have limited interpretations due to publication bias and confounding present in their source materials. In contrast to case reports, N-of-1 trials are prospective study designs that allow clinicians to make strong inferences about the effect of an intervention in a particular patient. They are double-blinded, single-patient, multicrossover studies that are of particular value in fields where it is difficult to recruit sufficient patients for conventional randomized control trials (RCTs), such as zoological medicine. Because they require multiple crossover periods, N-of-1 trials are ideal for evaluating short-acting interventions in patients with somewhat stable chronic diseases, such as osteoarthritis. More complex than conventional therapeutic trials, N-of-1 trials require prior consideration of how to achieve blinding, appropriate placebo controls, quantitative primary outcomes, analysis methods, and ethical approval. Aggregation of N-of-1 trials allows estimation of the average treatment effect across the population with fewer participants than a conventional RCT. Although systematic reviews and meta-analyses of case reports can be used to synthesize the observational N-of-1 data already in existence, N-of-1 trials offer an exciting way to prospectively generate strong evidence that will be useful for evidence-based decision-making.

Using Single-Case Experimental Design and Patient-Reported Outcome Measures to Evaluate the Treatment of Cancer-Related Cognitive Impairment in Clinical Practice

Cancer-related cognitive impairment (CRCI) affects a large proportion of cancer survivors and has significant negative effects on survivor function and quality of life (QOL). Treatments for CRCI are being developed and evaluated. Memory and attention adaptation training (MAAT) is a cognitive-behavioral therapy (CBT) demonstrated to improve CRCI symptoms and QOL in previous research. The aim of this article is to describe a single-case experimental design (SCED) approach to evaluate interventions for CRCI in clinical practice with patient-reported outcome measures (PROs). We illustrate the use of contemporary SCED methods as a means of evaluating MAAT, or any CRCI treatment, once clinically deployed. With the anticipated growth of cancer survivorship and concurrent growth in the number of survivors with CRCI, the treatment implementation and evaluation methods described here can be one way to assess and continually improve CRCI rehabilitative services.

Comparison of Bayesian Networks, G-estimation and linear models to estimate causal treatment effects in aggregated N-of-1 trials with carry-over effects

BACKGROUND

The aggregation of a series of N-of-1 trials presents an innovative and efficient study design, as an alternative to traditional randomized clinical trials. Challenges for the statistical analysis arise when there is carry-over or complex dependencies of the treatment effect of interest.

METHODS

In this study, we evaluate and compare methods for the analysis of aggregated N-of-1 trials in different scenarios with carry-over and complex dependencies of treatment effects on covariates. For this, we simulate data of a series of N-of-1 trials for Chronic Nonspecific Low Back Pain based on assumed causal relationships parameterized by directed acyclic graphs. In addition to existing statistical methods such as regression models, Bayesian Networks, and G-estimation, we introduce a carry-over adjusted parametric model (COAPM).

RESULTS

The results show that all evaluated existing models have a good performance when there is no carry-over and no treatment dependence. When there is carry-over, COAPM yields unbiased and more efficient estimates while all other methods show some bias in the estimation. When there is known treatment dependence, all approaches that are capable to model it yield unbiased estimates. Finally, the efficiency of all methods decreases slightly when there are missing values, and the bias in the estimates can also increase.

CONCLUSIONS

This study presents a systematic evaluation of existing and novel approaches for the statistical analysis of a series of N-of-1 trials. We derive practical recommendations which methods may be best in which scenarios.

Adaptive Deep Brain Stimulation: From Experimental Evidence Toward Practical Implementation

Closed-loop adaptive deep brain stimulation (aDBS) can deliver individualized therapy at an unprecedented temporal precision for neurological disorders. This has the potential to lead to a breakthrough in neurotechnology, but the translation to clinical practice remains a significant challenge. Via bidirectional implantable brain-computer-interfaces that have become commercially available, aDBS can now sense and selectively modulate pathophysiological brain circuit activity. Pilot studies investigating different aDBS control strategies showed promising results, but the short experimental study designs have not yet supported individualized analyses of patient-specific factors in biomarker and therapeutic response dynamics. Notwithstanding the clear theoretical advantages of a patient-tailored approach, these new stimulation possibilities open a vast and mostly unexplored parameter space, leading to practical hurdles in the implementation and development of clinical trials. Therefore, a thorough understanding of the neurophysiological and neurotechnological aspects related to aDBS is crucial to develop evidence-based treatment regimens for clinical practice. Therapeutic success of aDBS will depend on the integrated development of strategies for feedback signal identification, artifact mitigation, signal processing, and control policy adjustment, for precise stimulation delivery tailored to individual patients. The present review introduces the reader to the neurophysiological foundation of aDBS for Parkinson's disease (PD) and other network disorders, explains currently available aDBS control policies, and highlights practical pitfalls and difficulties to be addressed in the upcoming years. Finally, it highlights the importance of interdisciplinary clinical neurotechnological research within and across DBS centers, toward an individualized patient-centered approach to invasive brain stimulation. © 2023 The Authors. Movement Disorders published by Wiley Periodicals LLC on behalf of International Parkinson and Movement Disorder Society.

Genetics-informed precision treatment formulation in schizophrenia and bipolar disorder

Genetically informed drug development and repurposing is an attractive prospect for improving patient outcomes in psychiatry; however, the effectiveness of these endeavors is confounded by heterogeneity. We propose an approach that links interventions implicated by disorder-associated genetic risk, at the population level, to a framework that can target these compounds to individuals. Specifically, results from genome-wide association studies are integrated with expression data to prioritize individual "directional anchor" genes for which the predicted risk-increasing direction of expression could be counteracted by an existing drug. While these compounds represent plausible therapeutic candidates, they are not likely to be equally efficacious for all individuals. To account for this heterogeneity, we constructed polygenic scores restricted to variants annotated to the network of genes that interact with each directional anchor gene. These metrics, which we call a pharmagenic enrichment score (PES), identify individuals with a higher burden of genetic risk, localized in biological processes related to the candidate drug target, to inform precision drug repurposing. We used this approach to investigate schizophrenia and bipolar disorder and reveal several compounds targeting specific directional anchor genes that could be plausibly repurposed. These genetic risk scores, mapped to the networks associated with target genes, revealed biological insights that cannot be observed in undifferentiated genome-wide polygenic risk score (PRS). For example, an enrichment of these partitioned scores in schizophrenia cases with otherwise low PRS. In summary, genetic risk could be used more specifically to direct drug repurposing candidates that target particular genes implicated in psychiatric and other complex disorders.

Changing Oar Rotation Axis Position Increases Catch Angle During Indoor and In-Field Para-Rowing: A Randomized Crossover Trial Verified by a Repeated Measurement Trial

A long rowing stroke length is crucial for adequate rowing performance. Therefore, the relocation of the oar from traditional “in front” (NORM) to “behind the rotation axis” (GATE) may increase (para) rowing performance. Thus, 15 able-bodied rowers (21.4 ± 3.6 years; 187 ± 8 cm; 85.4 ± 8.2 kg) completed indoor TANK rowing 2 min TimeTrials (2 min-TT) of GATE and NORM in a randomized order. Additionally, one elite Paralympic oarsman (37 years, 185 cm, 67 kg) performed a multiple single case in-field BOAT testing (24x2min-TT of GATE and NORM in a randomized order). GATE revealed significantly larger catch angles during TANK (+97.1 ± 120.4%; p = 0.001, SMD = 0.84) and BOAT (+11.9 ± 3.2%; p < 0.021; SMD = 2.69; Tau-U = 0.70) compared to NORM. While total stroke length, rowing power, and work per stroke increased in GATE during TANK (p < 0.010, SMD > 0.634), no such significant changes of these performance parameters between GATE and NORM were observed during BOAT (p > 0.021; SMD < 0.58; Tau-U < 0.29). Rowing economy-related parameters (power or speed per oxygen uptake) and boat speed also showed no significant differences between GATE und NORM during BOAT (p > 0.61; SMD < 0.31; Tau-U < 0.19). The shape of the force–angle curve (position of peak force and ratio between average and maximal force) remained unaffected from GATE during both TANK (p > 0.73, SMD < 0.1) and BOAT (p > 0.63; SMD < 0.60; Tau-U < 0.27). In conclusion, GATE shifted the entire rowing stroke towards the catch (+6.6 ± 1.8°) without notably affecting relevant performance parameters during BOAT. Particularly during crew rowing, the minimization of detrimental boat movements for perfect synchrony should be aimed for. Accordingly, the combined application of GATE and NORM (for different athletes in crew boats) may be beneficial for rowing synchronization.

Temporal dynamics of depression, cognitive performance and sleep in older persons with depressive symptoms and cognitive impairments: a series of eight single-subject studies

OBJECTIVES

To investigate the presence, nature and direction of the daily temporal association between depressive symptoms, cognitive performance and sleep in older individuals.

DESIGN, SETTING, PARTICIPANTS

Single-subject study design in eight older adults with cognitive impairments and depressive symptoms.

MEASUREMENTS

For 63 consecutive days, depressive symptoms, working memory performance and night-time sleep duration were daily assessed with an electronic diary and actigraphy. The temporal associations of depressive symptoms, working memory and total sleep time were evaluated for each participant separately with time-series analysis (vector autoregressive modeling).

RESULTS

For seven out of eight participants we found a temporal association between depressive symptoms and/or sleep and/or working memory performance. More depressive symptoms were preceded by longer sleep duration in one person (r = 0.39; p < .001), by longer or shorter sleep duration than usual in one other person (B = 0.49; p < .001), by worse working memory in one person (B = -0.45; p = .007), and by better working memory performance in one other person (B = 0.35; p = .009). Worse working memory performance was preceded by longer sleep duration (r = -.35; p = .005) in one person, by shorter or longer sleep duration in three other persons (B = -0.76; p = .005, B = -0.61; p < .001; B = -0.34; p = .002), and by more depressive symptoms in one person (B = -0.25; p = .009).

CONCLUSION

The presence, nature and direction of the temporal associations between depressive symptoms, cognitive performance and sleep differed between individuals. Knowledge of personal temporal associations may be valuable for the development of personalized intervention strategies in order to maintain their health, quality of life, functional outcomes and independence.

Single case designs for early phase behavioral translational research in health psychology

OBJECTIVE

The biomedical research community has long recognized that much of the basic research being conducted, whether in the biological, behavioral or social sciences, is not readily translated into clinical and public health applications. This translational gap is due in part to challenges inherent in moving research findings from basic or discovery research to applied research that addresses clinical or public health problems. In the behavioral and social sciences, research designs typically used in the early phases of translational research are small, underpowered "pilot" studies that may lack sufficient statistical power to test the research question of interest. While this approach is discouraged, these studies are often employed to estimate effect sizes before embarking on a larger trial with adequate statistical power to test the research hypothesis. The goal of this paper is to provide an alternative approach to early phase studies using single case designs (SCDs).

METHOD

Review basic principles of SCDs; provide a series of hypothetical SCD replication experiments to illustrate (1) how data from SCDs can be analyzed to test the effects of an intervention on behavioral and biological outcomes and (2) how sample sizes can be derived for larger randomized controlled trials (RCTs) based on clinically meaningful effects from SCDs; and review feedback between SCDs and RCTs.

RESULTS

The paper illustrates the use of SCD reversal and multiple baseline designs for early phase translational research.

CONCLUSION

SCDs provide a flexible and efficient platform for the use of experimental methods in early phase translational research. (PsycInfo Database Record (c) 2022 APA, all rights reserved).

Establishment of an International Collaborative Network for N-of-1 Trials and Single-Case Designs

In this article we briefly examine the unique features of Single-Case Designs (SCDs) (studies in a single participant), their history and current trends, and real-world clinical applications. The International Collaborative Network for N-of-1 Trials and Single-Case Designs (ICN) is a formal collaborative network for individuals with an interest in SCDs. The ICN was established in 2017 to support the SCD scientific community and provide opportunities for collaboration, a global communication channel, resource sharing and knowledge exchange. In May 2021, there were more than 420 members in 31 countries. A member survey was undertaken in 2019 to identify priorities for the ICN for the following few years. This article outlines the key priorities identified and the ICN's progress to date in these key areas including network activities (developing a communications strategy to increase awareness, collecting/sharing a comprehensive set of resources, guidelines and tips, and incorporating the consumer perspective) and scientific activities (writing position papers and guest editing special journal issues, exploring key stakeholder perspectives about SCDs, and working to streamline ethical approval processes for SCDs). The ICN provides a practical means to engage with this methodology through membership. We encourage clinicians, researchers, industry, and healthcare consumers to learn more about and conduct SCDs, and to join us in our mission of using SCDs to improve health outcomes for individuals and populations.

Guidance for the Harmonisation and Improvement of Economic Evaluations of Personalised Medicine

OBJECTIVE

The objective of this study was to develop guidance contributing to improved consistency and quality in economic evaluations of personalised medicine (PM), given current ambiguity about how to measure the value of PM as well as considerable variation in the methodology and reporting in economic evaluations of PM.

METHODS

A targeted literature review of methodological papers was performed for an overview of modelling challenges in PM. Expert interviews were held to discuss best modelling practice. A systematic literature review of economic evaluations of PM was conducted to gain insight into current modelling practice. The findings were synthesised and used to develop a set of draft recommendations. The draft recommendations were discussed at a stakeholder workshop and subsequently finalised.

RESULTS

Twenty-two methodological papers were identified. Some argued that the challenges in modelling PM can be addressed within existing methodological frameworks, others disagreed. Eighteen experts were interviewed. They believed large uncertainty to be a key concern. Out of 195 economic evaluations of PM identified, 56% addressed none of the identified modelling challenges. A set of 23 recommendations was developed. Eight recommendations focus on the modelling of test-treatment pathways. The use of non-randomised controlled trial data is discouraged but several recommendations are provided in case randomised controlled trial data are unavailable. The parameterisation of structural uncertainty is recommended. Other recommendations consider perspective and discounting; premature survival data; additional value elements; patient and clinician compliance; and managed entry agreements.

CONCLUSIONS

This study provides a comprehensive list of recommendations to modellers of PM and to evaluators and reviewers of PM models.

Is the N-of-1 method applicable in bodywork research? Lessons learned using a trial as a methodological pilot

N-of-1 trial designs have rarely been used in bodywork research. Using a recent trial as a methodological pilot, critical issues related to the applicability of N-of-1 trials to bodywork are discussed. These include the issues of carry-over effects, bias-controlling approaches and statistical analysis. The discussion highlights the importance of mixed methods and draws some suggestions for a future research program. N-of-1 trials could be used to provide insights about some essential elements of bodywork modalities and their effectiveness.

Correlation of damage score in PTOA with changes in stress on cartilage in an ovine model

Objective

There is a high risk of developing osteoarthritis (OA) following traumatic injury to the knee. Severe ligament injuries can disrupt the integrity of the multicomponent knee at both biological and biomechanical levels. We hypothesize changes in cartilages stresses could lead to tissue damage and development of OA.

Design

The in-vivo gait kinematics of the stifle (knee) joint of four adult female ovine subjects were recorded prior to and at ten-and-twenty weeks following partial ACL-MCL transection. The subjects were sacrificed and the experimental joint from each subject was mounted on a parallel robotic system programmed with the kinematic findings. Ten custom-built Fibre Bragg Grating optic sensors were arranged to measure contact stresses on the surface of the tibial plateau articular cartilage. These sensors provide the first accurate stress measurements in a joint during gait replication using the previously recorded in-vivo kinematics. The relationship between the results obtained and observed focal damage was assessed.

Results

The locations on the tibial plateaus that experienced the greatest change in contact stresses corresponded with the locations of focal damage development. No direct link was detected between individual animal differences in kinematics and variations in stress magnitudes or the development of focal cartilage damage.

Conclusions

The findings highlight the importance of mechanical stress determinants in the integrated set point for the knee (with individual variation), and how injury-related stress changes correlate with development of PTOA.

Artificial Intelligence and Hypertension: Recent Advances and Future Outlook

Prevention and treatment of hypertension (HTN) are a challenging public health problem. Recent evidence suggests that artificial intelligence (AI) has potential to be a promising tool for reducing the global burden of HTN, and furthering precision medicine related to cardiovascular (CV) diseases including HTN. Since AI can stimulate human thought processes and learning with complex algorithms and advanced computational power, AI can be applied to multimodal and big data, including genetics, epigenetics, proteomics, metabolomics, CV imaging, socioeconomic, behavioral, and environmental factors. AI demonstrates the ability to identify risk factors and phenotypes of HTN, predict the risk of incident HTN, diagnose HTN, estimate blood pressure (BP), develop novel cuffless methods for BP measurement, and comprehensively identify factors associated with treatment adherence and success. Moreover, AI has also been used to analyze data from major randomized controlled trials exploring different BP targets to uncover previously undescribed factors associated with CV outcomes. Therefore, AI-integrated HTN care has the potential to transform clinical practice by incorporating personalized prevention and treatment approaches, such as determining optimal and patient-specific BP goals, identifying the most effective antihypertensive medication regimen for an individual, and developing interventions targeting modifiable risk factors. Although the role of AI in HTN has been increasingly recognized over the past decade, it remains in its infancy, and future studies with big data analysis and N-of-1 study design are needed to further demonstrate the applicability of AI in HTN prevention and treatment.

Measuring the Internal Stress in Ovine Meniscus During Simulated In Vivo Gait Kinematics: A Novel Method Using Fibre Optic Technology

Changes in stress transferred across articular joints have been described as a substantial factor in the initiation and progression of joint disease such as post-traumatic osteoarthritis and have thus been of interest to biomechanical researchers. However, to date, stress magnitudes within the menisci have not been successfully measured. In this study, a novel method for measuring stress within the menisci is presented. Small Fibre Bragg Grating (FBG) sensors were inserted inside menisci and used to measure mechanical stress during replicated gait cycles. In-vitro stress measurements within the menisci were preformed for healthy gait and gait following surgical damage to the joints. Together with our capability to reproduce in vivo motions accurately, the improvements in fibre optic technology have allowed for the first direct measurement of mechanical stress in menisci.

Optimizing Aggregated N-Of-1 Trial Designs for Predictive Biomarker Validation: Statistical Methods and Theoretical Findings

Background and Significance: Parallel-group randomized controlled trials (PG-RCTs) are the gold standard for detecting differences in mean improvement across treatment conditions. However, PG-RCTs provide limited information about individuals, making them poorly optimized for quantifying the relationship of a biomarker measured at baseline with treatment response. In N-of-1 trials, an individual subject moves between treatment conditions to determine their specific response to each treatment. Aggregated N-of-1 trials analyze a cohort of such participants, and can be designed to optimize both statistical power and clinical or logistical constraints, such as allowing all participants to begin with an open-label stabilization phase to facilitate the enrollment of more acutely symptomatic participants. Here, we describe a set of statistical simulation studies comparing the power of four different trial designs to detect a relationship between a predictive biomarker measured at baseline and subjects' specific response to the PTSD pharmacotherapeutic agent prazosin.

Methods: Data was simulated from 4 trial designs: (1) open-label; (2) open-label + blinded discontinuation; (3) traditional crossover; and (4) open label + blinded discontinuation + brief crossover (the N-of-1 design). Designs were matched in length and assessments. The primary outcome, analyzed with a linear mixed effects model, was whether a statistically significant association between biomarker value and response to prazosin was detected with 5% Type I error. Simulations were repeated 1,000 times to determine power and bias, with varied parameters.

Results: Trial designs 2 & 4 had substantially higher power with fewer subjects than open label design. Trial design 4 also had higher power than trial design 2. Trial design 4 had slightly lower power than the traditional crossover design, although power declined much more rapidly as carryover was introduced.

Conclusions: These results suggest that an aggregated N-of-1 trial design beginning with an open label titration phase may provide superior power over open label or open label and blinded discontinuation designs, and similar power to a traditional crossover design, in detecting an association between a predictive biomarker and the clinical response to the PTSD pharmacotherapeutic prazosin. This is achieved while allowing all participants to spend the first 8 weeks of the trial on open-label active treatment.

Applying Machine Learning to Daily-Life Data From the TrackYourTinnitus Mobile Health Crowdsensing Platform to Predict the Mobile Operating System Used With High Accuracy: Longitudinal Observational Study

BACKGROUND

Tinnitus is often described as the phantom perception of a sound and is experienced by 5.1% to 42.7% of the population worldwide, at least once during their lifetime. The symptoms often reduce the patient's quality of life. The TrackYourTinnitus (TYT) mobile health (mHealth) crowdsensing platform was developed for two operating systems (OS)-Android and iOS-to help patients demystify the daily moment-to-moment variations of their tinnitus symptoms. In all platforms developed for more than one OS, it is important to investigate whether the crowdsensed data predicts the OS that was used in order to understand the degree to which the OS is a confounder that is necessary to consider.

OBJECTIVE

In this study, we explored whether the mobile OS-Android and iOS-used during user assessments can be predicted by the dynamic daily-life TYT data.

METHODS

TYT mainly applies the paradigms ecological momentary assessment (EMA) and mobile crowdsensing to collect dynamic EMA (EMA-D) daily-life data. The dynamic daily-life TYT data that were analyzed included eight questions as part of the EMA-D questionnaire. In this study, 518 TYT users were analyzed, who each completed at least 11 EMA-D questionnaires. Out of these, 221 were iOS users and 297 were Android users. The iOS users completed, in total, 14,708 EMA-D questionnaires; the number of EMA-D questionnaires completed by the Android users was randomly reduced to the same number to properly address the research question of the study. Machine learning methods-a feedforward neural network, a decision tree, a random forest classifier, and a support vector machine-were applied to address the research question.

RESULTS

Machine learning was able to predict the mobile OS used with an accuracy up to 78.94% based on the provided EMA-D questionnaires on the assessment level. In this context, the daily measurements regarding how users concentrate on the actual activity were particularly suitable for the prediction of the mobile OS used.

CONCLUSIONS

In the work at hand, two particular aspects have been revealed. First, machine learning can contribute to EMA-D data in the medical context. Second, based on the EMA-D data of TYT, we found that the accuracy in predicting the mobile OS used has several implications. Particularly, in clinical studies using mobile devices, the OS should be assessed as a covariate, as it might be a confounder.

Some t-tests for N-of-1 trials with serial correlation

N-of-1 trials allow inference between two treatments given to a single individual. Most often, clinical investigators analyze an individual's N-of-1 trial data with usual t-tests or simple nonparametric methods. These simple methods do not account for serial correlation in repeated observations coming from the individual. Existing methods accounting for serial correlation require simulation, multiple N-of-1 trials, or both. Here, we develop t-tests that account for serial correlation in a single individual. The development includes effect size and precision calculations, both of which are useful for study planning. We then use Monte Carlo simulation to evaluate statistical properties of these serial t-tests, namely, Type I and II errors, and confidence interval widths, and compare these statistical properties to those of analogous usual t-test. The serial t-tests clearly outperform the usual t-tests commonly used in reporting N-of-1 results. Examples from N-of-1 clinical trials in fibromyalgia patients and from a behavioral health setting exhibit how accounting for serial correlation can change inferences. These t-tests are easily implemented and more appropriate than simple methods commonly used; however, caution is needed when analyzing only a few observations.

Comparison of Aggregated N-of-1 Trials with Parallel and Crossover Randomized Controlled Trials Using Simulation Studies

Background: N-of-1 trials offer an innovative approach to delivering personalized clinical care together with population-level research. While increasingly used, these methods have raised some statistical concerns in the healthcare community. Methods: We discuss concerns of selection bias, carryover effects from treatment, and trial data analysis conceptually, then rigorously evaluate concerns of effect sizes, power and sample size through simulation study. Four variance structures for patient heterogeneity and model error are considered in a series of 5000 simulated trials with 3 cycles, which compare aggregated N-of-1 trials to parallel randomized controlled trials (RCTs) and crossover trials. Results: Aggregated N-of-1 trials outperformed both traditional parallel RCT and crossover designs when these trial designs were simulated in terms of power and required sample size to obtain a given power. N-of-1 designs resulted in a higher type-I error probability than parallel RCT and cross over designs when moderate-to-strong carryover effects were not considered or in the presence of modeled selection bias. However, N-of-1 designs allowed better estimation of patient-level random effects. These results reinforce the need to account for these factors when planning N-of-1 trials. Conclusion: N-of-1 trial designs offer a rigorous method for advancing personalized medicine and healthcare with the potential to minimize costs and resources. Interventions can be tested with adequate power with far fewer patients than traditional RCT and crossover designs. Operating characteristics compare favorably to both traditional RCT and crossover designs.

Power and Design Issues in Crossover-Based N-Of-1 Clinical Trials with Fixed Data Collection Periods

“N-of-1,” or single subject, clinical trials seek to determine if an intervention strategy is more efficacious for an individual than an alternative based on an objective, empirical, and controlled study. The design of such trials is typically rooted in a simple crossover strategy with multiple intervention response evaluation periods. The effect of serial correlation between measurements, the number of evaluation periods, the use of washout periods, heteroscedasticity (i.e., unequal variances among responses to the interventions) and intervention-associated carry-over phenomena on the power of such studies is crucially important for putting the yield and feasibility of N-of-1 trial designs into context. We evaluated the effect of these phenomena on the power of different designs for N-of-1 trials using analytical theory based on standard likelihood principles assuming an autoregressive lag 1, i.e., AR(1), serial correlation structure among the measurements as well as simulation studies. By evaluating the power to detect effects in many different settings, we show that the influence of serial correlation and heteroscedasticity on power can be substantial, but can also be mitigated to some degree through the use of appropriate multiple evaluation periods. We also show that the detection of certain types of carry-over effects can be heavily influenced by design considerations as well.