Adaptive Clinical Trial Designs with Surrogates: When Should We Bother?

The success of a new drug is assessed within a clinical trial using a primary endpoint, which is typically the true outcome of interest—for example, overall survival. However, regulators sometimes approve drugs using a surrogate outcome—an intermediate indicator that is faster or easier to measure than the true outcome of interest—for example, progression-free survival—as the primary endpoint when there is demonstrable medical need. Although using a surrogate outcome (instead of the true outcome) as the primary endpoint can substantially speed up clinical trials and lower costs, it can also result in poor drug-approval decisions because the surrogate is not a perfect predictor of the true outcome. In this paper, we propose combining data from both surrogate and true outcomes to improve decision making within a late-phase clinical trial. In contrast to broadly used clinical trial designs that rely on a single primary endpoint, we propose a Bayesian adaptive clinical trial design that simultaneously leverages both observed outcomes to inform trial decisions. We perform comparative statics on the relative benefit of our approach, illustrating the types of diseases and surrogates for which our proposed design is particularly advantageous. Finally, we illustrate our proposed design on metastatic breast cancer. We use a large-scale clinical trial database to construct a Bayesian prior and simulate our design on a subset of clinical trials. We estimate that our design would yield a 16% decrease in trial costs relative to existing clinical trial designs, while maintaining the same Type I/II error rates. This paper was accepted by J. George Shanthikumar for the Special Issue on Data-Driven Prescriptive Analytics.

Predicting with Proxies: Transfer Learning in High Dimension

Predictive analytics is increasingly used to guide decision making in many applications. However, in practice, we often have limited data on the true predictive task of interest and must instead rely on more abundant data on a closely related proxy predictive task. For example, e-commerce platforms use abundant customer click data (proxy) to make product recommendations rather than the relatively sparse customer purchase data (true outcome of interest); alternatively, hospitals often rely on medical risk scores trained on a different patient population (proxy) rather than their own patient population (true cohort of interest) to assign interventions. Yet, not accounting for the bias in the proxy can lead to suboptimal decisions. Using real data sets, we find that this bias can often be captured by a sparse function of the features. Thus, we propose a novel two-step estimator that uses techniques from high-dimensional statistics to efficiently combine a large amount of proxy data and a small amount of true data. We prove upper bounds on the error of our proposed estimator and lower bounds on several heuristics used by data scientists; in particular, our proposed estimator can achieve the same accuracy with exponentially less true data (in the number of features d). Finally, we demonstrate the effectiveness of our approach on e-commerce and healthcare data sets; in both cases, we achieve significantly better predictive accuracy as well as managerial insights into the nature of the bias in the proxy data. This paper was accepted by George Shanthikumar, big data and analytics.

Rapid, accurate, nucleobase detection using FnCas9

Rapid detection of pathogenic sequences or variants in DNA and RNA through a point-of-care diagnostic approach is valuable for accelerated clinical prognosis as has been witnessed during the recent COVID-19 outbreak. Traditional methods relying on qPCR or sequencing are difficult to implement in settings with limited resources necessitating the development of accurate alternative testing strategies that perform robustly. Here, we present FnCas9 Editor Linked Uniform Detection Assay (FELUDA) that employs a direct Cas9 based enzymatic readout for detecting nucleotide sequences and identifying nucleobase identity without the requirement of trans-cleavage activity of reporter molecules. We demonstrate that FELUDA is 100% accurate in detecting single nucleotide variants (SNVs) including heterozygous carriers of a mutation and present a simple design strategy in the form of a web-tool, JATAYU, for its implementation. FELUDA is semi quantitative, can be adapted to multiple signal detection platforms and can be quickly designed and deployed for versatile applications such as infectious disease outbreaks like COVID-19. Using a lateral flow readout within 1h, FELUDA shows 100% sensitivity and 97% specificity across all range of viral loads in clinical samples. In combination with RT-RPA and a smartphone application True Outcome Predicted via Strip Evaluation (TOPSE), we present a prototype for FELUDA for CoV-2 detection at home.

Decentralized Common Knowledge Oracles

We define and analyze three mechanisms for getting common knowledge, a posteriori truths about the world, onto a blockchain in a decentralized setting. We show that, when a reasonable economic condition is met, these mechanisms are individually rational, incentive compatible, and decide the true outcome of valid oracle queries in both the non-cooperative and cooperative settings. These mechanisms are based upon repeated games with two classes of players: queriers who desire to get common knowledge truths onto the blockchain and a pool of reporters who posses such common knowledge. Presented with a new oracle query, reporters have an opportunity to report the truth in return for a fee provided by the querier. During subsequent oracle queries, the querier has an opportunity to punish any reporters who did not report truthfully during previous rounds. While the set of reporters has the power to cause the oracle to lie, they are incentivized not to do so.

Tree-weighting for multi-study ensemble learners

Multi-study learning uses multiple training studies, separately trains classifiers on individual studies, and then forms ensembles with weights rewarding members with better cross-study prediction ability. This article considers novel weighting approaches for constructing tree-based ensemble learners in this setting. Using Random Forests as a single-study learner, we perform a comparison of either weighting each forest to form the ensemble, or extracting the individual trees trained by each Random Forest and weighting them directly. We consider weighting approaches that reward cross-study replicability within the training set. We find that incorporating multiple layers of ensembling in the training process increases the robustness of the resulting predictor. Furthermore, we explore the mechanisms by which the ensembling weights correspond to the internal structure of trees to shed light on the important features in determining the relationship between the Random Forests algorithm and the true outcome model. Finally, we apply our approach to genomic datasets and show that our method improves upon the basic multi-study learning paradigm.

Outcomes of total knee arthroplasty in arthritis using functional knee score

<p class="abstract"><strong>Background:</strong> Knee being a major weight bearing joint, arthritis of the knee is a common problem. Over time various conservative measures have been used to alleviate the arthritic pain, but complete pain relief has rarely been achieved. Total Knee Arthroplasty has an established place in the treatment of knee arthritis and is an effective surgical modality that provides immediate pain relief and enhances quality of life. Though most patients were satisfied by the immediate outcome of the surgery, further detailed interrogation revealed concern and inability to perform activities they previously used to do, thus arising the need for an objective method to measure the true outcome.</p><p class="abstract"><strong>Methods:</strong> There are very few established objective scoring methods to evaluate the outcome following a Total Knee Arthroplasty. In our study, we have used the ‘Functional Knee Score’ for the 30 patients with arthritis who underwent TKA using the midvastus approach under tourniquet cover.<strong></strong></p><p class="abstract"><strong>Results:</strong> The majority of the patients were from the age group of 61-70 years which accounts for 36.7% of patients in our study. The mean functional new knee society score preoperatively was 39.90 standard deviation of 3.055 which improved to 80.77 with standard deviation of 6.263 postoperatively, p&lt;0.001.</p><p class="abstract"><strong>Conclusions:</strong> Using knee society functional score, 10 patients had excellent results, 16 had a good result, 4 patients a had fair result.</p>

LOST TO FOLLOW-UP : REASONS AND TRUE OUTCOME FOR PATIENTS ON ARV TREATMENT IN TERATAI CLINIC BANDUNG INDONESIA

Klinik RSHS Teratai HIV telah merawat pasien dari September 2004 sampai Desember 2009, sebanyak 2.060 ODHA dan 1290 menerima ARV (63%). Selama periode ini, sebesar 11% (141 orang) kehilangan kontak lebih tinggi dari keseluruhan persentase di Indonesia yaitu 10,7% (Depkes, 2009). Tujuan dari penelitian ini adalah untuk menyelidiki hasil yang benar dan faktor hilang tindak lanjut ARV di klinik RSHS Teratai ARV. Penelitian ini menggunakan metode wawancara mendalam dengan 25 pasien HIV / AIDS bawah pengobatan ARV di klinik Teratai RSHS dan mangkir sejak tahun 2009 yang tinggal di daerah Bandung. Informan dipilih secara acak dan dihubungi untuk kunjungan rumah untuk diwawancarai oleh pewawancara yang berpengalaman. Peneliti melakukan wawancara mendalam di rumah pasien sekitar 60 menit. Data direkam oleh pita digital direkam, ditranskrip dan dicetak oleh dua peneliti bidang independen menggunakan analisis tematik.Hasil penelitian ini menunjukkan tujuh tema yang pengetahuan yang tidak benar, optimisme yang tidak realistis, masalah aadministratif, perasaan efek samping negatif, masalah interpersonal, masalah kecanduan, dan kurangnya perawatan diri. Oleh karena itu, penting untuk mengembangkan tindakan pencegahan melalui rincian kontak pasien, meminta pasien persetujuan untuk dihubungi, telepon, kunjungan rumah setelah 3 minggu kehilangan kontak, dan menyarankan pasien untuk mendaftar ke klinik ART lebih dekat dengan rumah mereka , dan mengembangkan sistem reminder untuk mengingatkan pasien untuk mengambil ARV tepat waktu. ABSTRACTTeratai clinic HIV RSHS has treated patients from September 2004 to December 2009, as much as 2060 PLWHA and 1290 receive ARV (63%). Over this period, 11% (141) were lost to follow-up that is higher than overall lost to follow up percentage in Indonesia which is 10.7% (MOH annually report December 2009). The aim of this study was to investigate the true outcomes and factors of lost follow up to ARV. This study used indepth interview method to 25 HIV/AIDS patients under ARV treatment in the Teratai clinic RSHS and lost to follow up since 2009 who lived in Bandung area. Informants were randomly selected and contacted for a home visit to be interviewed by experienced interviewer. The interviewers obtained verbal informed consent and conducted an in depth interview in the patients home. Interview took about 60 minutes. Data were recorded by digital tape recorded, transcribed and scored by two field researchers independently using thematic analysis. The result of this study shows seven themes which are incorrect knowledge, unrealistic optimism, administrative problems, feeling of negative side effect, interpersonal problem, addiction problem, and lack of self-care. Therefore, it is important to develop preventive measure through up-dating patients’ contact details, ask patients consent to be contacted, up date phone contact, home visit after 3 weeks lost of contact, advice patients to register to ART clinic closer to their homes, and develop a reminding system to remind patients to take ARV on time.