Interim monitoring of sequential multiple assignment randomized trials using partial information

The sequential multiple assignment randomized trial (SMART) is the gold standard trial design to generate data for the evaluation of multistage treatment regimes. As with conventional (single-stage) randomized clinical trials, interim monitoring allows early stopping; however, there are few methods for principled interim analysis in SMARTs. Because SMARTs involve multiple stages of treatment, a key challenge is that not all enrolled participants will have progressed through all treatment stages at the time of an interim analysis. Wu et al. (2021) propose basing interim analyses on an estimator for the mean outcome under a given regime that uses data only from participants who have completed all treatment stages. We propose an estimator for the mean outcome under a given regime that gains efficiency by using partial information from enrolled participants regardless of their progression through treatment stages. Using the asymptotic distribution of this estimator, we derive associated Pocock and O'Brien-Fleming testing procedures for early stopping. In simulation experiments, the estimator controls type I error and achieves nominal power while reducing expected sample size relative to the method of Wu et al. (2021). We present an illustrative application of the proposed estimator based on a recent SMART evaluating behavioral pain interventions for breast cancer patients.

Effectiveness of the SMART Sex Ed program among 13-18 year old English and Spanish speaking adolescent men who have sex with men

Adolescent men who have sex with men (AMSM) have a high HIV incidence and low utilization of testing and prevention services. However, very few HIV prevention programs exist that focus on the unique sexual health needs of AMSM. SMART is a stepped care package of eHealth interventions that comprehensively address the sexual and HIV prevention needs of AMSM. This study examines the impact of the first step of SMART, "SMART Sex Ed," on 13- to 18-year-old AMSM (n = 983) from baseline to three-month follow-up across 18 separate outcomes measuring HIV prevention attitudes, skills, and behaviors. We observed significant change from baseline to three-month post-intervention in nine HIV-related outcomes (e.g., receipt of HIV and STI test, HIV knowledge), as well as largely consistent effects across demographic subgroups (e.g., race, age, rural, low SES). Analyses observed no effects on condom use behaviors. SMART Sex Ed shows promise as an effective sexual health education program for diverse AMSM.

A SMART approach to optimizing delivery of an mHealth intervention among cancer survivors with posttraumatic stress symptoms

BACKGROUND/AIMS

Many cancer survivors who received intensive treatment such as hematopoietic stem cell transplantation (HCT) experience posttraumatic stress disorder (PTSD) symptoms. PTSD is associated with lower quality of life and other symptoms that require clinical treatment. The iterative treatment decisions that happen in clinical practice are not adequately represented in traditional randomized controlled trials (RCT) of PTSD treatments. The proposed stepped-care SMART design allows for evaluation of initial response to the Cancer Distress Coach mobile app; adaptive stepped-care interventions; and precision treatment strategies that tailor treatment selection to patient characteristics.

METHODS/DESIGN

HCT survivors (N = 400) reporting PTSD symptoms are being recruited at two cancer centers and randomly assigned to: 1) Cancer Distress Coach app or 2) Usual Care. The app includes educational and cognitive behavioral therapy (CBT)-based activities. Four weeks post-randomization, participants re-rate their PTSD symptoms and, based on intervention response, non-responders are re-randomized to receive video-conferenced sessions with a therapist: 3) coaching sessions in using the mobile app; or 4) CBT specific to HCT survivors. Participants complete outcome measures of PTSD, depression, and anxiety after Months 1, 3, and 6. Participant characteristics moderating intervention responses will be examined.

CONCLUSIONS

This novel adaptive trial design will afford evidence that furthers knowledge about optimizing PTSD interventions for HCT survivors. To our knowledge, this study is the first SMART design evaluating PTSD symptom management in cancer survivors. If successful, it could be used to optimize treatment among a range of cancer and other trauma survivors.

Novel approach to modeling high-frequency activity data to assess therapeutic effects of analgesics in chronic pain conditions

Osteoarthritis (OA) is a chronic condition often associated with pain, affecting approximately fourteen percent of the population, and increasing in prevalence. A globally aging population have made treating OA-associated pain as well as maintaining mobility and activity a public health priority. OA affects all mammals, and the use of spontaneous animal models is one promising approach for improving translational pain research and the development of effective treatment strategies. Accelerometers are a common tool for collecting high-frequency activity data on animals to study the effects of treatment on pain related activity patterns. There has recently been increasing interest in their use to understand treatment effects in human pain conditions. However, activity patterns vary widely across subjects; furthermore, the effects of treatment may manifest in higher or lower activity counts or in subtler ways like changes in the frequency of certain types of activities. We use a zero inflated Poisson hidden semi-Markov model to characterize activity patterns and subsequently derive estimators of the treatment effect in terms of changes in activity levels or frequency of activity type. We demonstrate the application of our model, and its advance over traditional analysis methods, using data from a naturally occurring feline OA-associated pain model.

Ascertaining properties of weighting in the estimation of optimal treatment regimes under monotone missingness

Dynamic treatment regimes operationalize precision medicine as a sequence of decision rules, one per stage of clinical intervention, that map up-to-date patient information to a recommended intervention. An optimal treatment regime maximizes the mean utility when applied to the population of interest. Methods for estimating an optimal treatment regime assume the data to be fully observed, which rarely occurs in practice. A common approach is to first use multiple imputation and then pool the estimators across imputed datasets. However, this approach requires estimating the joint distribution of patient trajectories, which can be high-dimensional, especially when there are multiple stages of intervention. We examine the application of inverse probability weighted estimating equations as an alternative to multiple imputation in the context of monotonic missingness. This approach applies to a broad class of estimators of an optimal treatment regime including both Q-learning and a generalization of outcome weighted learning. We establish consistency under mild regularity conditions and demonstrate its advantages in finite samples using a series of simulation experiments and an application to a schizophrenia study.

Results of the Brief Jail Mental Health Screen Across Repeated Jail Bookings

OBJECTIVE

The Brief Jail Mental Health Screen (BJMHS) is widely used at intake in county jails to identify detainees who may have serious mental illness and who should be referred for further mental health evaluation. The BJMHS may be administered multiple times across repeated jail bookings; however, the extent to which results may change over time is unclear. To that end, the authors examined the odds of screening positive on the BJMHS across repeated jail bookings.

METHODS

Data were drawn from the administrative and medical records of a large, urban county jail that used the BJMHS at jail booking. The study sample comprised BJMHS results for the 12,531 jail detainees who were booked at least twice during the 3.5-year period (N=41,965 bookings). Multilevel logistic modeling was used to examine changes over time overall and within the four decision rules (current psychiatric medication, prior hospitalization, two or more current symptoms, and referral for any other reason).

RESULTS

Results show that the odds of a positive screen overall increased with each jail booking, as did the odds of referral for any other reason. In contrast, the odds of screening positive for two or more current symptoms and prior hospitalization decreased. There was no change in the odds of screening positive for current psychiatric medication across bookings.

CONCLUSIONS

Findings show that BJMHS results changed across bookings. Further research is needed to determine whether changes reflect true changes in mental health status, issues with fidelity, the repeated nature of the screening process, or other factors.

Creating an mHealth App for Colorectal Cancer Screening: User-Centered Design Approach

Background

Patients are increasingly using mobile health (mHealth) apps to monitor their health and educate themselves about medical issues. Despite the increasing popularity of such apps, poor design and usability often lead to suboptimal continued use of these apps and subsequently to poor adherence to the behavior changes at which they are aimed. One solution to these design problems is for app developers to use user-centered design (UCD) principles to consider the context and needs of users during the development process.

Objective

This study aimed to present a case study on the design and development process for an mHealth app that uses virtual human technology (VHT) to encourage colorectal cancer (CRC) screening among patients aged 50 years and above.

Methods

We have first provided an overview of the project and discussed its utilization of VHT. We have then reviewed UCD principles and how they can be incorporated into the development of health apps. We have described how we used UCD processes during the app’s development. We have then discussed the unique roles played by communication researchers, computer scientists, clinicians, and community participants in creating an mHealth app that is credible, usable, effective, and accessible to its target audience.

Results

The principles of UCD were woven throughout the project development, with researchers collecting feedback from patients and providers at all stages and using that feedback to improve the credibility, usability, effectiveness, and accessibility of the mHealth app. The app was designed in an iterative process, which encouraged feedback and improvement of the app and allowed teams from different fields to revisit topics and troubleshoot problems.

Conclusions

Implementing a UCD process contributed to the development of an app, which not only reflected cross-disciplinary expertise but also the needs, wants, and concerns of patients.

Time course and prognostic value of serum GFAP, pNFH, and S100β concentrations in dogs with complete spinal cord injury because of intervertebral disc extrusion

Background

A noninvasive biomarker is needed to predict recovery from severe spinal cord injury (SCI) because of thoracolumbar intervertebral disc extrusion (TL‐IVDE). Proteins released from neural and glial cells can be detected in the blood and show promise as prognostic tools, but their concentration is influenced by time after injury.

Hypothesis/Objectives

Serum concentrations of glial fibrillary acidic protein (GFAP), phosphorylated neurofilament heavy chain (pNFH), and S100β will follow different time courses; measurement of combinations of these proteins will predict outcome.

Animals

Thirty‐one dogs with TL‐IVDE causing paralysis with no pain perception.

Methods

Prospective study. Serum samples were taken at presentation and intervals over 56 days and banked at −80°C. Glial fibrillary acidic protein, pNFH, and S100β concentrations were measured using ELISA tests and plotted against time from onset of nonambulatory status. Outcome was established at 6 months. The association between biomarker concentration and outcome was examined using logistic regression, receiver operator characteristics curve analysis, and model development.

Results

Thirty‐one dogs participated, 3/31 (10%) developed progressive myelomalacia and 19/31 (62%) recovered ambulation. Glial fibrillary acidic protein and S100β concentrations rose for the first 1 to 3 days, and were undetectable by 14 and 28 days, respectively. Phosphorylated neurofilament heavy chain concentrations peaked at 14 days and were detectable at 56 days. Glial fibrillary acidic protein concentrations in the first 72 hours after onset of nonambulatory status predicted recovery with an accuracy of 76.7%‐89% depending on sample timing.

Conclusions and Clinical Importance

Serum GFAP concentrations can be used to predict outcome in clinically complete SCI. A rapid inexpensive bedside test is needed.

Predictors of Response to 4-Aminopyridine in Chronic Canine Spinal Cord Injury

4-Aminopyridine (4AP), a potassium channel antagonist, can improve hindlimb motor function in dogs with chronic thoracolumbar spinal cord injury (SCI); however, individual response is variable. We hypothesized that injury characteristics would differ between dogs that do and do not respond to 4AP. Our objective was to compare clinical, electrodiagnostic, gait, and imaging variables between dogs that do and do not respond to 4AP, to identify predictors of response. Thirty-four dogs with permanent deficits after acute thoracolumbar SCI were enrolled. Spasticity, motor and sensory evoked potentials (MEPs, SEPs), H-reflex, F-waves, gait scores, and magnetic resonance imaging (MRI) with diffusion tensor imaging (DTI) were evaluated at baseline and after 4AP administration. Baseline variables were assessed as predictors of response; response was defined as ≥1 point change in open field gait score. Variables were compared pre- and post-4AP to evaluate 4AP effects. Fifteen of 33 (45%) dogs were responders, 18/33 (55%) were non-responders and 1 was eliminated because of an adverse event. Pre-H-reflex threshold <1.2 mA predicted non-response; pre-H-reflex threshold >1.2 mA and Canine Spasticity Scale overall score <7 were predictive of response. All responders had translesional connections on DTI. MEPs were more common post-4AP than pre-4AP (10 vs. 6 dogs) and 4AP decreased H-reflex threshold and increased spasticity in responders. 4-AP impacts central conduction and motor neuron pool excitability in dogs with chronic SCI. Severity of spasticity and H-reflex threshold might allow prediction of response. Further exploration of electrodiagnostic and imaging characteristics might elucidate additional factors contributing to response or non-response.

Quantifying center of pressure variability in chondrodystrophoid dogs

The center of pressure (COP) position reflects a combination of proprioceptive, motor and mechanical function. As such, it can be used to quantify and characterize neurologic dysfunction. The aim of this study was to describe and quantify the movement of COP and its variability in healthy chondrodystrophoid dogs while walking to provide a baseline for comparison to dogs with spinal cord injury due to acute intervertebral disc herniations. Fifteen healthy adult chondrodystrophoid dogs were walked on an instrumented treadmill that recorded the location of each dog's COP as it walked. Center of pressure (COP) was referenced from an anatomical marker on the dogs' back. The root mean squared (RMS) values of changes in COP location in the sagittal (y) and horizontal (x) directions were calculated to determine the range of COP variability. Three dogs would not walk on the treadmill. One dog was too small to collect interpretable data. From the remaining 11 dogs, 206 trials were analyzed. Mean RMS for change in COPx per trial was 0.0138 (standard deviation, SD 0.0047) and for COPy was 0.0185 (SD 0.0071). Walking speed but not limb length had a significant effect on COP RMS. Repeat measurements in six dogs had high test retest consistency in the x and fair consistency in the y direction. In conclusion, COP variability can be measured consistently in dogs, and a range of COP variability for normal chondrodystrophoid dogs has been determined to provide a baseline for future studies on dogs with spinal cord injury.

Associations Between Anesthetic Variables and Functional Outcome in Dogs With Thoracolumbar Intervertebral Disk Extrusion Undergoing Decompressive Hemilaminectomy

Background

Outcome of acute experimental spinal cord injury is strongly associated with tissue perfusion and oxygenation. Cardiopulmonary depression could affect outcome in dogs undergoing general anesthesia for surgical treatment of thoracolumbar intervertebral disk extrusion (IVDE).

Hypothesis/Objectives

To evaluate the effects of general anesthesia on functional outcome in dogs undergoing surgery to treat thoracolumbar IVDE.

Animals

Eighty‐four client‐owned dogs with acute thoracolumbar IVDE treated by decompressive hemilaminectomy.

Methods

Exploratory, retrospective observational study. Medical records were reviewed for clinical presentation and anesthetic monitoring variables, including duration of anesthesia and surgery, hypotension, bradycardia, temperature, and respiratory parameters. Multivariable regression tree analysis was performed to explore associations between anesthetic variables and functional outcome scores after 6 weeks, as well as return to ambulatory status.

Results

Episodes of bradycardia (69%) and hypotension (57%) were frequent. Across all outcome measures, regression tree analysis highlighted functional grade at presentation as the primary determining factor, and among pain perception negative dogs, there was a possible association between increased duration of surgery and poorer outcome. In dogs with intact pain perception, duration of bradycardia, mean body temperature, and mean end‐tidal carbon dioxide were highlighted.

Conclusions and Clinical Importance

Exploratory statistical methods can facilitate hypothesis‐generating studies to inform prospective investigations in veterinary medicine. Although the mechanism is uncertain, increased duration of surgery might be associated with poorer outcome in pain perception negative dogs with thoracolumbar IVDE.

On Sparse representation for Optimal Individualized Treatment Selection with Penalized Outcome Weighted Learning

As a new strategy for treatment which takes individual heterogeneity into consideration, personalized medicine is of growing interest. Discovering individualized treatment rules (ITRs) for patients who have heterogeneous responses to treatment is one of the important areas in developing personalized medicine. As more and more information per individual is being collected in clinical studies and not all of the information is relevant for treatment discovery, variable selection becomes increasingly important in discovering individualized treatment rules. In this article, we develop a variable selection method based on penalized outcome weighted learning through which an optimal treatment rule is considered as a classification problem where each subject is weighted proportional to his or her clinical outcome. We show that the resulting estimator of the treatment rule is consistent and establish variable selection consistency and the asymptotic distribution of the estimators. The performance of the proposed approach is demonstrated via simulation studies and an analysis of chronic depression data.

Potassium channel antagonists 4-aminopyridine and the T-butyl carbamate derivative of 4-aminopyridine improve hind limb function in chronically non-ambulatory dogs; a blinded, placebo-controlled trial

4-Aminopyridine (4-AP) blocks voltage gated potassium channels, restoring conduction to demyelinated axons and improving function in demyelinating conditions, but its use is associated with adverse effects and benefit in spinal cord injury is limited. Derivatives of 4-AP have been developed to improve clinical efficacy while reducing toxicity. We compared the therapeutic effects of orally administered 4-AP and its t-butyl carbamate derivative (t-butyl) with placebo in dogs that had suffered an acute spinal cord injury that left them chronically paralyzed. Nineteen dogs were entered into the trial, conducted in two-week treatment blocks starting with placebo, followed by random assignment to 4-AP or t-butyl, a washout and then the opposite medication followed by placebo. Investigators and owners were blinded to treatment group. Primary outcome measures included open field gait score (OFS), and treadmill based stepping score and regularity index, with additional secondary measures also considered. Thirteen of 19 dogs completed the protocol. Two were euthanized due to unrelated heath problems, two developed side effects and two were unable to complete for unrelated reasons. Dogs showed significant improvement in supported stepping score (from 17.39 to 37.24% with 4-AP; 16.85 to 29.18% with t-butyl p<0.0001) and OFS (from 3.63 to 4.73 with 4-AP; 3.78 to 4.45 with t-butyl, p = 0.005). Response was individually variable and most dramatic in three dogs that were able to walk without support with treatment. No significant difference was found between 4-AP and t-butyl. No adverse effects were reported with t-butyl but gastrointestinal upset and seizures were observed in two dogs with 4-AP. In conclusion, both 4-AP and t-butyl significantly improved supported stepping ability in dogs with chronic spinal cord injury with no adverse effects noted with t-butyl. Drug response varied widely between individuals, highlighting the need to understand the factors that influence canine and human patients' response to therapy.

Personalized Evaluation of Biomarker Value: A Cost-Benefit Perspective

For a patient who is facing a treatment decision, the added value of information provided by a biomarker depends on the individual patient's expected response to treatment with and without the biomarker, as well as his/her tolerance of disease and treatment harm. However, individualized estimators of the value of a biomarker are lacking. We propose a new graphical tool named the subject-specific expected benefit curve for quantifying the personalized value of a biomarker in aiding a treatment decision. We develop semiparametric estimators for two general settings: (i) when biomarker data are available from a randomized trial; and (ii) when biomarker data are available from a cohort or a cross-sectional study, together with external information about a multiplicative treatment effect. We also develop adaptive bootstrap confidence intervals for consistent inference in the presence of nonregularity. The proposed method is used to evaluate the individualized value of the serum creatinine marker in informing treatment decisions for the prevention of renal artery stenosis.

A multiple imputation strategy for sequential multiple assignment randomized trials

Sequential multiple assignment randomized trials (SMARTs) are increasingly being used to inform clinical and intervention science. In a SMART, each patient is repeatedly randomized over time. Each randomization occurs at a critical decision point in the treatment course. These critical decision points often correspond to milestones in the disease process or other changes in a patient's health status. Thus, the timing and number of randomizations may vary across patients and depend on evolving patient-specific information. This presents unique challenges when analyzing data from a SMART in the presence of missing data. This paper presents the first comprehensive discussion of missing data issues typical of SMART studies: we describe five specific challenges and propose a flexible imputation strategy to facilitate valid statistical estimation and inference using incomplete data from a SMART. To illustrate these contributions, we consider data from the Clinical Antipsychotic Trial of Intervention and Effectiveness, one of the most well-known SMARTs to date.

Estimating Optimal Treatment Regimes from a Classification Perspective

A treatment regime maps observed patient characteristics to a recommended treatment. Recent technological advances have increased the quality, accessibility, and volume of patient-level data; consequently, there is a growing need for powerful and flexible estimators of an optimal treatment regime that can be used with either observational or randomized clinical trial data. We propose a novel and general framework that transforms the problem of estimating an optimal treatment regime into a classification problem wherein the optimal classifier corresponds to the optimal treatment regime. We show that commonly employed parametric and semi-parametric regression estimators, as well as recently proposed robust estimators of an optimal treatment regime can be represented as special cases within our framework. Furthermore, our approach allows any classification procedure that can accommodate case weights to be used without modification to estimate an optimal treatment regime. This introduces a wealth of new and powerful learning algorithms for use in estimating treatment regimes. We illustrate our approach using data from a breast cancer clinical trial.

Nifedipine vasodilates human forearm arteries and dorsal hand veins constricted by specific alpha-adrenoceptor stimulation

1. The local dilative effect of the calcium entry blocker nifedipine on forearm arteries and dorsal hand veins has been studied in 27 healthy male volunteers. 2. Nifedipine induced an increase of blood flow by 1190% (P less than 0.001) in the forearm. 3. The construction of the hand veins induced by stimulation of either postsynaptic alpha 1- or alpha 2-adrenoceptors was reduced (P less than 0.001) by nifedipine. 4. The calcium entry blocker nifedipine is a potent dilator of human forearm arteries as well as of dorsal hand veins.

Specific alpha-adrenoceptor-mediated vasoconstriction in human veins and interaction with the calcium entry blockers nifedipine and diltiazem

In order to characterize the pharmacological properties of post-junctional alpha-adrenoceptors, the vasoconstricting effects of the non-specific alpha-adrenoceptor agonist noradrenaline, of the specific alpha 1-adrenoceptor agonist phenylephrine and of the specific alpha 2-adrenoceptor agonist azepexole and the interaction with the calcium entry blockers nifedipine and diltiazem were studied in hand veins of eight healthy volunteers. Both subtypes of alpha-adrenoceptors could be demonstrated in human hand veins in vivo. The alpha 1-subtype seemed to prevail. Vasoconstriction induced by stimulation of either post-synaptic alpha 1- or alpha 2-adrenoceptors was reduced by the calcium entry blockers diltiazem and nifedipine.