Recruitment Hypotheses Research Articles

Enhancing site selection strategies in clinical trial recruitment using real-world data modeling.

Slow patient enrollment or failing to enroll the required number of patients is a disruptor of clinical trial timelines. To meet the planned trial recruitment, site selection strategies are used during clinical trial planning to identify research sites that are most likely to recruit a sufficiently high number of subjects within trial timelines. We developed a machine learning approach that outperforms baseline methods to rank research sites based on their expected recruitment in future studies. Indication level historical recruitment and real-world data are used in the machine learning approach to predict patient enrollment at site level. We define covariates based on published recruitment hypotheses and examine the effect of these covariates in predicting patient enrollment. We compare model performance of a linear and a non-linear machine learning model with common industry baselines that are constructed from historical recruitment data. Performance of the methodology is evaluated and reported for two disease indications, inflammatory bowel disease and multiple myeloma, both of which are actively being pursued in clinical development. We validate recruitment hypotheses by reviewing the covariates relationship with patient recruitment. For both indications, the non-linear model significantly outperforms the baselines and the linear model on the test set. In this paper, we present a machine learning approach to site selection that incorporates site-level recruitment and real-world patient data. The model ranks research sites by predicting the number of recruited patients and our results suggest that the model can improve site ranking compared to common industry baselines.

Revisiting sardine recruitment hypotheses: Egg‐production‐based survival index improves understanding of recruitment mechanisms of fish under climate variability

AbstractProportionality between spawning stock biomass (SSB) and annual total egg production (TEP) has been an inherent assumption in spawner–recruitment models for fisheries management and studies on recruitment mechanisms in fish. However, recent evidence of density dependence in egg production has led to a need to reconsider the validity of recruitment per spawning stock biomass (RPS), which is a SSB‐based survival index from hatching to recruitment, in recruitment studies. Here we revisit sardine recruitment hypotheses based on recruitment per egg production (RPE), which is a TEP‐based survival index, by reanalyzing published data for the Pacific stock of Japanese sardine (Sardinops melanostictus) in the Kuroshio Current system at a multidecadal scale. By adopting RPE, we detected statistically significant relationships of the survival from hatching to recruitment to large‐scale ocean climate anomaly, ambient temperature in the nursery grounds and growth rate during the early life stages. None of these relationships were detected by RPS. We propose the adoption of TEP‐based survival indices rather than SSB‐based ones in future studies on recruitment mechanisms of fish in relation to environmental and biological factors. Developing egg‐production‐based survival indices has the potential to improve understanding of recruitment mechanisms of fish under climate variability.

Investigating environmental influence and temporal changes in sole (Solea solea) larvae condition using histology

In the eastern part of the English Channel, common sole (Solea solea) has strong interests in fisheries research. Low recruitment along with a decline in spawning stock biomass have been observed for several years. According to the recruitment hypotheses, larval survival may play an important role that needs to be considered. The fish larval condition can be assessed using histology which has been recognised as the most appropriate method to provide a reliable index of the nutritional status. Based on this approach, this study aimed to identify critical periods of wild-collected sole larvae and to determine sources of variations of their condition between two periods separated by more than 20 years. In line with other studies, the transition from endogenous to exogenous feeding was identified as the most critical period with the lowest proportion of healthy larvae observed. During this first feeding stage, good larval conditions were located in sampling stations close to the coast and at the end of the spring season, in relation to higher temperature and fluorescence values. This highlights the need for sole larvae survival to cross the coastal front, which splits the central and coastal waters, to reach more stable and productive areas. This coastal migration pattern was consistent between 1995 and 2017, with however significantly lower larval abundances in the recent period. Multivariate analyses showed that the spring environmental conditions of 1995, characterised with lower temperature and higher fluorescence values were more favourable to larval condition, compared to the spring in 2017. Areas providing suitable environmental conditions in 2017 were more restricted and limited to sampling stations in front of estuaries. Since small differences in larval survival can lead to large fluctuations in recruitment, the larval condition should be studied in a more long-term approach. This would provide a better understanding of the environmental influence on larval survival and recruitment success.

A fine-scale multi-step approach to understand fish recruitment variability

Recruitment is one of the dominant processes regulating fish population productivity. It is, however, notoriously difficult to predict, as it is the result of a complex multi-step process. Various fine-scale drivers might act on the pathway from adult population characteristics to spawning behaviour and egg production, and then to recruitment. Here, we provide a holistic analysis of the Northwest Atlantic mackerel recruitment process from 1982 to 2017 and exemplify why broad-scale recruitment–environment relationships could become unstable over time. Various demographic and environmental drivers had a synergetic effect on recruitment, but larval survival through a spatio-temporal match with prey was shown to be the key process. Recruitment was also mediated by maternal effects and a parent–offspring fitness trade-off due to the different feeding regimes of adults and larvae. A mismatch curtails the effects of high larval prey densities, so that despite the abundance of food in recent years, recruitment was relatively low and the pre-existing relationship with overall prey abundance broke down. Our results reaffirm major recruitment hypotheses and demonstrate the importance of fine-scale processes along the recruitment pathway, helping to improve recruitment predictions and potentially fisheries management.

Riverscape recruitment: a conceptual synthesis of drivers of fish recruitment in rivers

Most fish recruitment models consider only one or a few drivers in isolation, rarely include species’ traits, and have limited relevance to riverine environments. Despite their diversity, riverine fishes share sufficient characteristics that prediction of recruitment should be possible. Here we synthesize the essential components of fish recruitment hypotheses and the key features of rivers to develop a model that predicts relative recruitment strength, for all fishes, in rivers under all flow conditions. The model proposes that interactions between flow and physical complexity will create locations in rivers, at mesoscales, where energy and nutrients are enriched. The resultant production of small prey will be concentrated and prey and fish larvae located (through dispersal or retention) so that the larvae can feed, grow, and recruit. Our synthesis explains how flow and physical complexity affect fish recruitment and provides a conceptual basis to better conserve and manage riverine fishes globally.

İhtiyoplanktonun dağılım dinamikleri ve stoka katılım hipotezleri

Planktonic eggs and larvae of marine teleost fishes are considered as ichthyoplankton until they improve active locomotion. These fishes usually produce a large number of propagules however, the great majority of them die before recruitment due to the extrinsic conditions such as predation, starvation and aberrant drift. In this period, small variations in survival rates may cause important differences in recruitment success. On the other hand, the presence of the ideal habitats which can sustain maximal survival for propagules, can significantly change in space and time. Therefore, it is critically important that emergence of propagules should match with the appearance of ideal habitats. To ensure this match, fishes have advanced different spawning strategies. In this review, intrinsic (spawning strategies) and extrinsic (environmental factors) conditions affecting the spatio-temporal distribution of ichthyoplankton have been evaluated on a theoretical basis by discussing match-mismatch dynamics and recruitment hypotheses.

Mechanisms driving recruitment variability in fish: comparisons between the Laurentian Great Lakes and marine systems

AbstractIn his seminal work, Hjort (in Fluctuations in the great fisheries of Northern Europe. Conseil Parmanent International Pour L'Exploration De La Mar. Rapports et Proces-Verbaux, 20: 1–228, 1914) observed that fish population levels fluctuated widely, year-class strength was set early in life, and egg production by adults could not alone explain variability in year-class strength. These observations laid the foundation for hypotheses on mechanisms driving recruitment variability in marine systems. More recently, researchers have sought to explain year-class strength of important fish in the Laurentian Great Lakes and some of the hypotheses developed for marine fisheries have been transferred to Great Lakes fish. We conducted a literature review to determine the applicability of marine recruitment hypotheses to Great Lakes fish. We found that temperature, interspecific interactions, and spawner effects (abundance, age, and condition of adults) were the most important factors in explaining recruitment variability in Great Lakes fish, whereas relatively fewer studies identified bottom-up trophodynamic factors or hydrodynamic factors as important. Next, we compared recruitment between Great Lakes and Baltic Sea fish populations and found no statistical difference in factors driving recruitment between the two systems, indicating that recruitment hypotheses may often be transferable between Great Lakes and marine systems. Many recruitment hypotheses developed for marine fish have yet to be applied to Great Lakes fish. We suggest that future research on recruitment in the Great Lakes should focus on forecasting the effects of climate change and invasive species. Further, because the Great Lakes are smaller and more enclosed than marine systems, and have abundant fishery-independent data, they are excellent candidates for future hypothesis testing on recruitment in fish.

Unquantifiable uncertainty in projecting stock response to climate change: Example from North East Arctic cod

Data suggest that for some years there has been a positive relationship between the recruitment of cod in the Barents Sea and the sea temperature at the Kola section during the year of spawning. However, analysis of the most recent data indicates that this relationship no longer holds. This change in the recruitment dynamics will clearly have an impact on our understanding of future stock dynamics and long-term yield. It also highlights the impacts arising from possible future changes in similar relationships in other species and ecosystems on our ability to predict biological responses to climate change arising. This article uses a ‘STOCOBAR’ forward simulation model to evaluate North East Arctic cod dynamics under a variety of climate scenarios and recruitment hypotheses, presenting the differences in modelled spawning stock biomass under temperature-dependent and -independent recruitment situations. The divergence between the modelled populations (and hence yields) under the different recruitment hypotheses indicates the high difficulty of predicting the future development of a stock with any degree of certainty, or even with any quantifiable degree of uncertainty. These results highlight the importance of having a management regime that is robust to unpredicted and unpredictable changes in stock dynamics, and the need for management strategy evaluations under a wide range of possible future scenarios.

Recruitment processes in Baltic sprat – A re-evaluation of GLOBEC Germany hypotheses

The GLOBEC Germany program (2002–2007) had the ambitious goal to resolve the processes impacting the recruitment dynamics of Baltic sprat (Sprattus sprattus L.) by examining various factors affecting early life history stages. At the start of the research program, a number of general recruitment hypotheses were formulated, i.e. focusing on (1) predation, (2) food availability, (3) physical parameters, (4) the impact of current systems, and finally (5) the importance of top-down vs bottom-up effects. The present study synthesizes the results of field sampling (2002 and 2003), laboratory experiments, and modeling studies to re-evaluate these hypotheses for the Baltic sprat stock. Recruitment success was quite different in the 2years investigated. Despite a lower spawning stock biomass in 2003, the total number of recruits was almost 2-fold higher that year compared to 2002. The higher recruitment success in 2003 could be attributed to enhanced survival success during the post-larval/juvenile stage, a life phase that appears to be critical for recruitment dynamics. In the state of the Baltic ecosystem during the period of investigation, we consider bottom-up control (e.g. temperature, prey abundance) to be more important than top-down control (predation mortality). This ranking in importance does not vary seasonally. Prevailing water circulation patterns and the transport dynamics of larval cohorts have a strong influence on sprat recruitment success. Pronounced transport to coastal areas is detrimental for year-class strength particularly at high sprat stock sizes. A suggested mechanism is density-dependant regulation of survival via intra- and inter-specific competition for prey in coastal areas. A documented change in larval vertical migration behavior between the early 1990s and early 2000s increased the transport potential to the coast, strengthening the coupling between inter-annual differences in the magnitude and direction of wind-driven surface currents and year-to-year changes in reproductive success. However, due to the strong linkages and feed-back loops in the Baltic Sea food web, the most robust projections of the future strength of the Baltic sprat stock will need to take into account climate-driven changes in both abiotic (e.g., drift trajectories) and biotic (trophodynamic) factors. Although our understanding of processes affecting pre-recruit (larval) growth and survival has been advanced by the integrated research conducted within the GLOBEC Germany program, key mechanisms potentially affecting life stages outside of the spawning basins remain to be explored including the dynamics of coastal habitats of juveniles and the feeding and overwintering grounds of adults.

Predation dynamics of mackerel on larval and juvenile anchovy: is capture success linked to prey condition?

We tested whether the predation dynamics of chub mackerel Scomber japonicus and spotted mackerel S. australasicus on young anchovy Engraulis japonicus relates to individual growth characteristics of the prey and could account for the growth-selective survival predicted by recruitment hypotheses. Juvenile and adult mackerel were sampled along with their young anchovy prey field in 2004 (juvenile mackerel and larval anchovy) and 2005 (adult mackerel and juvenile anchovy) off the Pacific coast of Honshu, Japan. The recent 5-day mean growth rate of larval and juvenile survivors and prey found in the stomach of mackerel was estimated from the otolith microstructure. No significant difference was found between the recent growth of larval or juvenile survivors and that of preyed individuals. We conclude that despite a relatively small body size, the high activity level and predation skills displayed by mackerel prevent fast-growing larvae and early juveniles from benefitting in terms of the expected survival advantage over slow-growers. Hence, growth-selective predation mortality of larval fish would depend on the feeding ecology of the predator rather than predator size. Selection for fast growth is more likely to occur under predation pressure from invertebrate organisms and small pelagic fish specialized on zooplankton, such as herring and anchovy.

Asian Risk Seeking and Overconfidence1

Weber & Hsee (1998) proposed the “cushion effect” as an explanation of the higher risk taking observed in some East Asian samples, relative to Western samples. Yates, Lee, & Shinotsuka (1992) proposed that the overconfidence observed in most Asian countries, relative to Western countries, reflects differences in the number of arguments typically recruited in those countries. This study examined the 2 seemingly separated domains of research: risk taking and overconfidence. It compared individual and collective (both family and group) decisions among Chinese in Singapore. This permitted tests of both the cushion and the argument recruitment hypotheses. The overall results obtained no support for the cushion effect and partial support for the argument recruitment hypotheses.

American Kestrel reproduction: evidence for the selection hypothesis and the role of dispersal

Many hypotheses attempt to explain why younger, less experienced birds have a relatively low reproductive output. We evaluated reproductive patterns of marked American Kestrels Falco sparverius nesting in boxes in southwestern Idaho from 1992 to 2006 to test predictions of these hypotheses. Results were consistent with the selection (differential mortality) hypothesis and did not support the constraint, restraint or recruitment hypotheses. Most known‐age Kestrels nested in their first year of life, and there was no apparent short‐term or long‐term reproductive advantage to delayed breeding. The number of years that Kestrels nested in study area boxes ranged from 1 to 6 years, with most Kestrels nesting in only 1 year. Reproductive rates were higher for birds with at least 1 year of nesting experience than for birds nesting in boxes for the first time. After 2 years of nesting, reproductive rates levelled off; there was no evidence for additional improvement or for senescence. Differences in reproductive output with experience/age were due to variation among and not within individuals. Individuals that nested in more than 1 year had similar reproductive rates in their first and second years. Poor producers either died or dispersed after 1 year of nesting in study area boxes. Successful females that nested early in the season and successful males that had been produced locally had the highest probability of returning to nest in a subsequent year.

Exploring the Generality of Recruitment Hypotheses for Largemouth Bass along a Latitudinal Gradient of Florida Lakes

AbstractLatitudinal influences on growing season length and winter severity cause variation in prerecruitment life history across species distributions. We evaluated two recruitment hypotheses for a broadly distributed species, largemouth bass Micropterus salmoides, at the southern extent of their natural range. We tested (1) whether early hatching provided a growth and survival advantage relative to later hatching through their first summer and (2) whether overwinter size‐selective mortality strongly influenced recruitment to age 1 across Florida's latitudinal gradient. We sampled the 2003 and 2004 year‐classes at six Florida lakes that spanned latitudes from 27°0′N to 30°5′N. Our results did not fully conform to our hypotheses or the results frequently reported from more northerly latitudes because largemouth bass that hatched early did not exhibit a growth and survival advantage at all lakes and we did not detect strong size‐selective overwinter mortality. Early hatching at south Florida lakes resulted in slow growth for three of four lake and year combinations and early‐hatched fish never exhibited a survival advantage relative to later‐hatched fish, which was probably due to cool water temperatures soon after hatching. Environmental variability influenced the interactions between hatching period, growth, and survival for Florida's largemouth bass populations, which has implications for understanding fish recruitment across broad spatial scales.

Detecting critical periods in larval flatfish populations

We evaluate the time-course of deaths and evidence of periods of increased mortality (i.e., critical periods) in laboratory populations of larval flatfish. First, we make the distinction between age-at-death and abundance-at-time data for fish larvae, the latter being typical in studies of natural populations. Next, we describe an experimental investigation of age- and temperature-dependent mortality in larval winter flounder, Pseudopleuronectes americanus. The survivorship curves of these populations differed significantly in both the magnitude and time-course of mortality among the four water temperatures evaluated (7, 10, 13, and 16°C). Mortality was highest in the cooler temperatures and concentrated in the third quarter of larval life, largely concurrent with settlement of surviving members of the cohort. Among the statistical methods for analysing survival data, the proportional-hazards model with time-varying covariates proved best at capturing the patterns of age-specific mortalities. We conclude that fair appraisals of recruitment hypotheses which are predicated on periods of high, age-specific mortality that vary with environmental conditions (e.g., Hjort's critical period hypothesis) will require: (1) data that are based on age, not time; (2) data that are of higher temporal resolution than commonly available at present and (3) analytical methods that are sensitive to irregularities in survivorship curves. We suggest four research approaches for evaluating critical periods in nature.

Dancing bees and the language controversy

For 30 years a controversy has raged over whether the honey bee waggle dance communicates the distance, direction, and scent of a food source to other bees, or whether it communicates only the scent. One way or the other the dance functions to recruit other bees to the food source visited by the dancing bee. Both sides of the debate have disputed the interpretation of results presented to date, but more data, rather than more arguing, are required to resolve the controversy. Two experiments are presented that test predictions of the two leading recruitment hypotheses: Karl von Frisch's dance language hypothesis, which suggests the dance is a symbolic language conveying directions to a food source, and Adrian Wenner's odor search hypothesis, which suggests the dance conveys no information other than the scent of a food source. The results indicate that recruits are indeed learning the direction of a food source when they follow dances, as von Frisch asserted 50 years ago.

Recruitment and recruitment processes of winter flounder, Pleuronectes americanus, at different latitudes, implications of an individual-based simulation model

We evaluate the joint effects of temperature, hours of daylight, prey production, spawning regimes, and mortality rates on recruitment of winter flounder, Pleuronectes americanus, by means of an individual-based simulation model. These factors were varied to approximate conditions at three latitudinally distinct sites in winter flounder's geographic range from New Jersey, USA (southernmost site) to New Brunswick, Canada (northernmost site). Most life historical and demographic response variables from simulation output differed substantially among sites. At year's end, the recruits were larger but fewer at lower latitudes than at higher ones. We further assessed the effects of season duration, prey production, and mortality rate on recruitment by means of a 3 3-factorial numerical experiment. This analysis provided estimates of the proportion of variance in the response variables (the number and sizes of recruits, larval and juvenile growth rates, larval period duration) due to each of the three manipulated factors and their interactions. Season duration explained most of the variation in the responses except for the sizes of recruits at year's end which was influenced most by the timing of prey production. Interaction effects of season duration and prey production were evident in all responses, reflecting the importance to flounder recruitment of the phenology of prey production relative to flounder early life history. Even though mortality rates were assigned as part of the factorial design structure, they accounted for only 10% of the variance in the number of recruits. These simulation results are in need of empirical verification and we suggest ways to proceed. We also propose means of enhancing data on winter flounder life history and ecology, and methods for expanding our model and analysis. We conclude that our approach proved useful for 1. appraising multiple, interacting recruitment processes, 2. locating areas where incomplete knowledge of winter flounder life history and the habitat it occupies limits a more satisfactory assessment of recruitment processes, and 3. identifying recruitment hypotheses that warrant further empirical evaluation.

An overview of the second flatfish symposium: Recruitment in flatfish

The Second International Flatfish Symposium was focused around recruitment and opened with a challenge and a question. The challenge was to intensify and clarify our thinking about recruitment. The question was: why, after nearly a century of concerted efforts, do we not yet understand (have the ability to predict) variability in year-class strength of fishes? Two answers were suggested: 1. we lack sufficient data and, 2. we do not yet have an adequate framework for interpreting existing data. The structuring of the symposium around 7 specific recruitment hypotheses provided a framework for presentations and discussion. Each hypothesis was reviewed by a keynote speaker followed by oral presentations and a poster session. Lastly, the information presented on each topic was summarized by the key-note speaker (Key) to lead to a discussion. In this paper, the conclusions on the various hypotheses will be discussed, then the suggestions for future research, and, finally, the need for data versus a new framework will be considered. In the course of the discussion, some other, more generic, problems in recruitment investigations that were underscored by the Symposium will be highlighted.

Recruitment, Search Behavior, and Flight Ranges of Honey Bees

During the past three decades, considerable evidence has been gathered in attempts to understand more fully honey bee recruitment to food sources. Those efforts also apply directly to two long-standing and competing recruitment hypotheses: odor search vs . “dance language” communication. However, whereas most researchers have focused on individual interactions and behavior, the colony can also be viewed as a unit. A review of evidence from a colony perspective reveals that colony members range an average distance from their home base, whether while foraging on food sources, while collecting water, or while relocating as swarms. Those averages, based on the logarithm of the distance from the colony, vary with the type of resource exploited and size of the odor field. Such a mathematical correspondence between distances travelled from parent colonies may well agree with an odorsearch recruitment model, but is hardly reconcilable with the “dance language” hypothesis.

Modeling of Recruitment Patterns in Pacific Cod (Gadus macrocephalus) in Hecate Strait, British Columbia

We combined published recruitment hypotheses into a single modeling framework and explored the relative contributions of the hypothesis factors to recruitment variation in Pacific cod (Gadus macrocephalus). Our five hypotheses were (1) recruitment decreased with increasing transport during the larval period; (2) recruitment was related to percentage hatching success as a dome-shaped curve with a peak between 4.0 and 5.0 °C; (3) recruitment was positively related to abundance of age-0 juvenile Pacific herring (Clupea harengus pallasi) hatched in the same year as the cod; (4) recruitment was positively related to abundance of age-0 and age-1 herring in the year prior to spawning; and (5) a dome-shaped curve related recruitment to adult stock size. We used new records for temperature, and a new index of advection based on measures of current in Hecate Strait. For fitting a response surface with age-3 cod recruitment as the independent variable, we used a 20-yr data series of variables from 1962 to 1981 inclusive and reserved data for the years 1982–85 for a prediction test. Only hypotheses 1 and 5 were supported by both the response surface fit and prediction tests. The stock–recruitment curve can be fitted only when the transport curve is fitted simultaneously.

Experimental Harvesting of Cyclic Stocks in the Face of Alternative Recruitment Hypotheses

Cyclic fluctuations in year-class abundance have been recorded for several fish populations. Historical patterns are generally consistent with several hypotheses about underlying mechanisms, which involve (a) strong density dependence or (b) environmental forcing. Optimal harvesting policies differ depending on which hypothesis is deemed true. Not only does the harvesting policy affect the amount of catch taken, but it can strongly affect our ability to discriminate between alternative hypotheses. Fully-adaptive feedback policies, which explicitly account for the information content of harvest controls, were computed with dynamic progamming for some model prototypes of the problem. Results indicate that deliberate experimentation with escapement levels, oriented to reduce the uncertainty, may be worthwhile under some conditions. In most cases, however, optimal passive policies (in which learning occurs passively) were a good approximation to actively adaptive policies. The optimal passive controls for maximizing a risk-neutral objective, such as total catch, were generally very informative. Although passive and active controls markedly differed in some cases, expected discounted yields under the two types of policies were very similar (differences less than 2%). The extreme risk-averse presciption of managing a stock according to the more pessimistic recruitment model may lead to considerable yield losses.