Evaluating individualized treatment effect predictions: A model-based perspective on discrimination and calibration assessment.

In recent years, there has been a growing interest in the prediction of individualized treatment effects. While there is a rapidly growing literature on the development of such models, there is little literature on the evaluation of their performance. In this paper, we aim to facilitate the validation of prediction models for individualized treatment effects. The estimands of interest are defined based on the potential outcomes framework, which facilitates a comparison of existing and novel measures. In particular, we examine existing measures of discrimination for benefit (variations of the c-for-benefit), and propose model-based extensions to the treatment effect setting for discrimination and calibration metrics that have a strong basis in outcome risk prediction. The main focus is on randomized trial data with binary endpoints and on models that provide individualized treatment effect predictions and potential outcome predictions. We use simulated data to provide insight into the characteristics of the examined discrimination and calibration statistics under consideration, and further illustrate all methods in a trial of acute ischemic stroke treatment. The results show that the proposed model-based statistics had the best characteristics in terms of bias and accuracy. While resampling methods adjusted for the optimism of performance estimates in the development data, they had a high variance across replications that limited their accuracy. Therefore, individualized treatment effect models are best validated in independent data. To aid implementation, a software implementation of the proposed methods was made available in R.

Regularized parametric survival modeling to improve risk prediction models.

We propose to combine the benefits of flexible parametric survival modeling and regularization to improve risk prediction modeling in the context of time-to-event data. Thereto, we introduce ridge, lasso, elastic net, and group lasso penalties for both log hazard and log cumulative hazard models. The log (cumulative) hazard in these models is represented by a flexible function of time that may depend on the covariates (i.e., covariate effects may be time-varying). We show that the optimization problem for the proposed models can be formulated as a convex optimization problem and provide a user-friendly R implementation for model fitting and penalty parameter selection based on cross-validation. Simulation study results show the advantage of regularization in terms of increased out-of-sample prediction accuracy and improved calibration and discrimination of predicted survival probabilities, especially when sample size was relatively small with respect to model complexity. An applied example illustrates the proposed methods. In summary, our work provides both a foundation for and an easily accessible implementation of regularized parametric survival modeling and suggests that it improves out-of-sample prediction performance.

The potential benefit of statin prescription based on prediction of treatment responsiveness in older individuals: An application to the PROSPER randomised controlled trial.

BACKGROUND & AIMS: Clinical guidelines often recommend to treat individuals based on their cardiovascular risk. We revisit this paradigm and quantify the efficacy of three treatment strategies: (i) overall prescription, i.e. treatment to all individuals sharing the eligibility criteria of a trial; (ii) risk-stratified prescription, i.e. treatment only to those at an elevated outcome risk; and (iii) prescription based on predicted treatment responsiveness. METHODS: We reanalysed the PROSPER randomised controlled trial, which included individuals aged 70-82 years with a history of, or risk factors for, vascular diseases. We conducted the derivation and internal-external validation of a model predicting treatment responsiveness. We compared to placebo (n= 2913): (i) pravastatin (n= 2891); (ii) pravastatin in the presence of previous vascular diseases and placebo in the absence thereof (n= 2925); and (iii) pravastatin in the presence of a favourable prediction of treatment response and placebo in the absence thereof (n= 2890). RESULTS: We found an absolute difference in primary outcome events composed of coronary death, non-fatal myocardial infarction, fatal or non-fatal stroke, per 10 000 person-years equal to: -78 events (95% CI, -144 to -12) when prescribing pravastatin to all participants; -66 events (95% CI, -114 to -18) when treating only individuals with an elevated vascular risk; and -103 events (95% CI, -162 to -44) when restricting pravastatin to individuals with a favourable prediction of treatment response. CONCLUSIONS: Pravastatin prescription based on predicted responsiveness may have an encouraging potential for cardiovascular prevention. Further external validation of our results and clinical experiments are needed.

This study invistigates whether an algorithm to predict how much old age individuals would benefit from a statin treatment could be useful to guide clinicians in their prescription decision-making; the key findings are: About one out of seven individuals included in the study has no predicted benefit of pravastatin; Compared to prescribing pravastatin to all old age individuals at risk of cardiovascular diseases, withholding pravastatin in those with no predicted benefit seems to lead to a better prevention of cardiovascular events.

High prevalence of peripheral neuropathy in multiple myeloma patients and the impact of vitamin D levels, a cross-sectional study.

PURPOSE: Peripheral neuropathy (PN) is common in patients with multiple myeloma (MM). We hypothesized that the relationship between hypovitaminosis D and PN described in diabetes mellitus patients may also be present in MM patients. METHODS: To study this potential association, we assessed the incidence of hypovitaminosis D (vitamin D < 75 nmol/L [= 30 ng/mL]) in smouldering and active MM patients in two Dutch hospitals. Furthermore, a validated questionnaire was used to distinguish different PN grades. RESULTS: Of the 120 patients included between January 2017 and August 2018, 84% had an inadequate vitamin D level (median vitamin D level 49.5 nmol/L [IQR 34-65 nmol/L]; mean age: 68 years [SD ± 7.7]; males: 58%). PN was reported by 69% of patients (n = 83); however, of these 83 patients, PN was not documented in the medical records of 52%. An association was found between lower vitamin D levels and higher incidence of PN in the total population (P = 0.035), and in the active MM patients (P = 0.016). CONCLUSION: This multi-centre cohort study showed that PN and hypovitaminosis D are common in MM patients, and addressing low vitamin D levels in the treatment of MM patients might be beneficial in reducing the risk of PN. More attention for PN is warranted, as PN is underreported by clinicians. Further research is needed to fully understand the implications of vitamin D in the development of PN in patients with MM. CLINICAL TRIAL REGISTRATION: Netherland Trial Register NL5835, date of registration July 28, 2016.

The use of terrestrial and aquatic microcosms and mesocosms for the ecological risk assessment of veterinary medicinal products.

In this paper, we investigate the applicability of experimental model ecosystems (microcosms and mesocosms) for the ecological risk assessment of veterinary medicinal products (VMPs). VMPs are used in large quantities, but the assessment of associated risks to the environment is limited, although they are continually infused into the environment via a number of routes. It is argued that the experience obtained by pesticide research largely can be used when evaluating VMPs, although there are several major differences between pesticides and pharmaceuticals (e.g., knowledge of their mechanisms of action on nontarget organisms). Also, because microorganisms are often the target organisms of VMPs, risk assessment should focus more on endpoints describing functional processes. This paper provides a review of the current risk assessment schemes of Europe and North America along with examples of experiments already performed with veterinary medicinal products in aquatic and terrestrial ecosystem models. We suggest that some of the approaches developed for pesticide risk assessment can be used for VMPs and offer suggestions for the development of a framework for ecological risk assessment of VMPs.

Professional autonomy and the normative structure of medical practice.

Professional autonomy is often described as a claim of professionals that has to serve primarily their own interests. However, it can also be seen as an element of a professional ideal that can function as a standard for professional, i.e. medical practice. This normative understanding of the medical profession and professional autonomy faces three threats today. 1) Internal erosion of professional autonomy due to a lack of internal quality control by the medical profession; 2) the increasing upward pressure on health care expenses that calls for a health care policy that could imply limitations for the professional autonomy of physicians; 3) a distorted understanding of the profession as being based on a formal type of knowledge and related technology, in which other normative dimensions of medical practice are neglected and which frustrates meaningful communication between physicians and patients. To answer these threats a normative structure analysis of medical practice is presented, that indicates which principles and norms are constitutive for medical practice. It is concluded that professional autonomy, normatively understood, should be maintained to avoid the lure of the technological imperative and to protect patients against third parties' pressure to undertreatment. However, this professional autonomy can only be maintained if members of the profession subject their activities and decisions to a critical evaluation by other members of the profession and by patients and if they continue to critically reflect on the values that regulate today's medicine.

Population genetics meets behavioral ecology.

Populations are often composed of more than just randomly mating subpopulations - many organisms from social groups with distinct patterns of mating and dispersal. Such patterns have recieved much attention in behavioral ecology, yet theories of population genetics rarely take social structures into account. Consequently, population geneticists often report high levels of apparent in breeding and concomitantly low efective sizes, even for species that avoid mating between close kin. Recently, a view of gene dynamics has been introduced that takes dispersal and social structure into account. Accounting for social structure in population genetics leads to a different perspective on how genetic variation is partitoned and the rate at which genic diversity is lost in natural populations - a view that is more consistent with observed behaviors for the minimization of inbreeding.

Infanticide in prairie dogs: lactating females kill offspring of close kin.

Infanticide, although common in a wide range of species including humans and other primates, is poorly understood. A 7-year study under natural conditions reveals that infanticide within colonies of black-tailed prairie dogs (Cynomys ludovicianus) is striking for three reasons. It is the major source of juvenile mortality, accounting for the total or partial demise of 51 percent of all litters born. The most common killers are resident lactating females. The most common victims are the offspring of close kin.

Prairie dogs avoid extreme inbreeding.

Black-tailed prairie dogs (Rodentia: Sciuridae: Cynomys ludovicianus) live in colonies composed of contiguous but separate family groups called coteries. During the 6 years that individuals in a colony were observed, they almost nevermated with close genetic relatives. Inbreeding is avoided in four ways: (i) a young male usually leaves his natal coterie before breeding, but his female relatives remain; (ii) an adult male usually leaves his breeding coterie before his daughters mature; (iii) a young female is less likely to come into estrus if her father is in her coterie; and (iv) an estrous female behaviorally avoids mating with a father, son, or brother in her coterie.