Processing math: 100%
Research article Special Issues

Differences in how interventions coupled with effective reproduction numbers account for marked variations in COVID-19 epidemic outcomes

  • Received: 05 June 2020 Accepted: 21 July 2020 Published: 27 July 2020
  • The COVID-19 outbreak, designated a "pandemic" by the World Health Organization (WHO) on 11 March 2020, has spread worldwide rapidly. Each country implemented prevention and control strategies, mainly classified as SARS LCS (SARS-like containment strategy) or PAIN LMS (pandemic influenza-like mitigation strategy). The reasons for variation in each strategy's efficacy in controlling COVID-19 epidemics were unclear and are investigated in this paper. On the basis of the daily number of confirmed local (imported) cases and onset-to-confirmation distributions for local cases, we initially estimated the daily number of local (imported) illness onsets by a deconvolution method for mainland China, South Korea, Japan and Spain, and then estimated the effective reproduction numbers Rt by using a Bayesian method for each of the four countries. China and South Korea adopted a strict SARS LCS, to completely block the spread via lockdown, strict travel restrictions and by detection and isolation of patients, which led to persistent declines in effective reproduction numbers. In contrast, Japan and Spain adopted a typical PAIN LMS to mitigate the spread via maintaining social distance, self-quarantine and isolation etc., which reduced the Rt values but with oscillations around 1. The finding suggests that governments may need to consider multiple factors such as quantities of medical resources, the likely extent of the public's compliance to different intensities of intervention measures, and the economic situation to design the most appropriate policies to fight COVID-19 epidemics.

    Citation: Fan Xia, Yanni Xiao, Peiyu Liu, Robert A. Cheke, Xuanya Li. Differences in how interventions coupled with effective reproduction numbers account for marked variations in COVID-19 epidemic outcomes[J]. Mathematical Biosciences and Engineering, 2020, 17(5): 5085-5098. doi: 10.3934/mbe.2020274

    Related Papers:

    [1] Marie Vandekerckhove, Yu-lin Wang . Emotion, emotion regulation and sleep: An intimate relationship. AIMS Neuroscience, 2018, 5(1): 1-17. doi: 10.3934/Neuroscience.2018.1.1
    [2] Daniel A. Barone, Ana C. Krieger . The Function of Sleep. AIMS Neuroscience, 2015, 2(2): 71-90. doi: 10.3934/Neuroscience.2015.2.71
    [3] Mark R. Zielinski, James T. McKenna, Robert W. McCarley . Functions and Mechanisms of Sleep. AIMS Neuroscience, 2016, 3(1): 67-104. doi: 10.3934/Neuroscience.2016.1.67
    [4] Carolyn E. Jones, Miranda M. Lim . Phasic Sleep Events Shape Cognitive Function after Traumatic Brain Injury: Implications for the Study of Sleep in Neurodevelopmental Disorders. AIMS Neuroscience, 2016, 3(2): 232-236. doi: 10.3934/Neuroscience.2016.2.232
    [5] Kimberly A Cote, Catherine E Milner, Tamara A Speth . Altered Sleep Mechanisms following Traumatic Brain Injury and Relation to Waking Function. AIMS Neuroscience, 2015, 2(4): 203-228. doi: 10.3934/Neuroscience.2015.4.203
    [6] Alexis M. Chambers, Jessica D. Payne . The Influence of Sleep on the Consolidation of Positive Emotional Memories: Preliminary Evidence. AIMS Neuroscience, 2014, 1(1): 39-51. doi: 10.3934/Neuroscience.2014.1.39
    [7] Monireh Asadi Ghaleni, Forouzan Fattahi Masrour, Narjes Saryar, Alexandra J. Bratty, Ebrahim Norouzi, Matheus Santos de Sousa Fernandes, Georgian Badicu . Effects of an intervention combining physical activity and components of Amygdala and Insula Retraining (AIR) on sleep and working memory among older male adults. AIMS Neuroscience, 2024, 11(4): 421-438. doi: 10.3934/Neuroscience.2024025
    [8] Alexis M. Chambers, Jessica D. Payne . Neural Plasticity and Learning: The Consequences of Sleep. AIMS Neuroscience, 2014, 1(2): 163-168. doi: 10.3934/Neuroscience.2014.2.163
    [9] Ritwik Das, Artur Luczak . Epileptic seizures and link to memory processes. AIMS Neuroscience, 2022, 9(1): 114-127. doi: 10.3934/Neuroscience.2022007
    [10] Eduardo Mercado III . Relating Cortical Wave Dynamics to Learning and Remembering. AIMS Neuroscience, 2014, 1(3): 185-209. doi: 10.3934/Neuroscience.2014.3.185
  • The COVID-19 outbreak, designated a "pandemic" by the World Health Organization (WHO) on 11 March 2020, has spread worldwide rapidly. Each country implemented prevention and control strategies, mainly classified as SARS LCS (SARS-like containment strategy) or PAIN LMS (pandemic influenza-like mitigation strategy). The reasons for variation in each strategy's efficacy in controlling COVID-19 epidemics were unclear and are investigated in this paper. On the basis of the daily number of confirmed local (imported) cases and onset-to-confirmation distributions for local cases, we initially estimated the daily number of local (imported) illness onsets by a deconvolution method for mainland China, South Korea, Japan and Spain, and then estimated the effective reproduction numbers Rt by using a Bayesian method for each of the four countries. China and South Korea adopted a strict SARS LCS, to completely block the spread via lockdown, strict travel restrictions and by detection and isolation of patients, which led to persistent declines in effective reproduction numbers. In contrast, Japan and Spain adopted a typical PAIN LMS to mitigate the spread via maintaining social distance, self-quarantine and isolation etc., which reduced the Rt values but with oscillations around 1. The finding suggests that governments may need to consider multiple factors such as quantities of medical resources, the likely extent of the public's compliance to different intensities of intervention measures, and the economic situation to design the most appropriate policies to fight COVID-19 epidemics.


    In many biomedical applications, the primary interest centers on predicting a survival outcome, for instance, the t-year survival probability, or the median survival time for future patients. For some diseases, it may be of much relevance to predict the survival function over a continuum of time for better treatment and surveillance. The problem of survival prediction is often tackled by first formulating a regression model that relates the survival time to the covariates and then making the prediction according to the fitted model. The commonly used approaches to assess the survival rate (or disease risk) are either based on modeling the association between the baseline covariates and the failure times (e.g. [1,2,3,4]) or through modeling the relationship between the hazard function and baseline covariates (e.g. [5,6,7]). As an alternative, the censored quantile regression [8,9,10] provides a valuable complement to the aforementioned methods. The censored quantile regression method has great advantages in the interpretation of regression coefficients which are derived under distribution-free assumptions. However, the censored quantile regression method focuses on a single quantile at a time, hence fails to make full use of the quantile information of the target distribution.

    The regression-based prediction directly models the conditional hazard function or the conditional regression function. The information of the covariates is incorporated and the resulting model can also be used for quantifying the risks for individual patients. However, the prediction accuracy of the regression approach relies heavily on whether the model is correctly specified. When a misspecified model is used, the prediction results can be misleading. However, in practice, it is often difficult to specify a correct model, especially when patients population are heterogeneous, or the data structure is complex. Furthermore, predicting the conditional survival outcome for individual patients is often too difficult or unrealistic. For example, Henderson and Keiding [11] convincingly showed that statistical models and indices can be useful at the group or population level, but may have limited predictive values for individual survival since human survival is so uncertain. Therefore throughout the paper, we focus on predicting unconditional survival outcomes. For such purposes, the conditional approach does not directly target the quantity to predict and is hence less ideal.

    So far, to the best of our knowledge, there is very limited research discussing the survival-rate (or disease-risk) assessment by matching quantiles or survival distributions. For complete data, the idea of matching quantiles is explored in many contexts (e.g. [12,13,14]). The matching quantiles estimation (MQE) method is proved to be an effective approach to assess the target distribution. For regression models, the MQE method shares certain similarities in form with the ordinary least squares estimation (OLS) and the quantile regression (QR) method [12,15]. But the MQE method is quite different from the classical methods such as the QR method. To be specific, the MQE is proposed to assess the (unconditional) target distribution, while the QR method is used for estimating regression coefficients based on conditional quantile functions. The MQE method makes use of both information of the order and the distance between quantiles of the target distribution and those used for matching.

    One advantage of the MQE method is that it can be implemented by matching the local quantiles between τ1 and τ2 only (0<τ1<τ2<1). This could be very attractive if we are only interested in studying a specific part of the target distribution, such as the middle or the lower end of the target distribution. Another advantage is that it does not require the observations being paired, i.e., the size of the sample from the target distribution and that of the counterpart are allowed to be unequal. It makes the MQE method more appealing and practical than traditional methods, especially for missing data. Sgouropoulos et al. [14] propose a MQE method by matching the sample quantiles of target distribution with that of a linear combination of covariates, which uses an iterative procedure based on permutation and OLS in computation. Although the iterative algorithm is fast, it inherits several disadvantages from OLS such as being sensitive to outliers, inapplicable to unpaired observations as well as the incomplete data due to censoring.

    Motivated by the MQE method, we propose a matching censored quantiles approach for predicting the survival rates and assessing the target distribution of interest. Particularly, the proposed method not only bears certain similarities with the classical quantile regression method and the composite quantile regression method [16], but also maintains major advantages of the aforementioned MQE methods for complete data. In addition, the proposed method avoids using permutations in the computational algorithm and can be easily extended to match a complex transformation of the target distribution. Last but not the least, the proposed method provides an alternative to assess or predict the target distribution in the presence of right-censored data.

    The rest of the article is organized as follows. In Section 2, we first present some notations, and then introduce the matching censored quantiles method. In Section 3, we provide the asymptotic properties of the proposed estimator. Section 4 discusses the matching measurement criteria. Section 5 presents extensive simulation studies. An illustrative example is provided in Section 6. Finally, Section 7 concludes with some remarks.

    Let Ti be the failure time of the i-th subject, and Ci be the censoring time. Denote the observed time as Yi=min(Ti,Ci), and the censoring indicator as Δi=I(TiCi). Let Zi=(1,Zi1,,Zip)T be a (p+1)×1 vector of covariates for the ith subject. The observations of {(Yi,Δi,Zi),i=1,,n} are independent and identically distributed copies of (Y,Δ,Z). The censoring mechanism is assumed to be non-informative, i.e., Ti and Ci are independent of each other, or Ti and Ci are conditionally independent of each other given Z.

    To assess the survival rates of T, we aim to find a transformation G(βTZ) such that its distribution matches the distribution of T as close as possible, where G() is a known, continuous and strictly increasing function, and βRp+1 is a (p+1)-dimensional coefficient. Denote the cumulative distribution function of T and βTZ as FT(t)=Pr(Tt) and FG(βTZ)(t)=Pr{G(βTZ)t}, respectively. Correspondingly, we write the survival function of T and βTZ as ST(t)=1FT(t) and SG(βTZ)(t)=1FG(βTZ)(t).

    Let H(t)=G1(t) be the inverse function of G(t), then H() is also a known, continuous and strictly increasing function. Note the fact that

    FT(t)=Pr(Tt)=Pr{H(T)H(t)}=FH(T){H(t)},FG(βTZ)(t)=Pr{G(βTZ)t}=Pr{βTZH(t)}=FβTZ{H(t)},

    hence, to search β such that FG(βTZ) matches FT is equivalent to find a linear combination βTZ such that its distribution matches the distribution of H(T).

    With complete data, Sgouropoulos et al. [14] proposed to use the distribution of a linear combination βTZ to match the target distribution, and β is estimated by minimizing the objective function,

    minβni=1{T(i)(βTZ)(i)}2, (2.1)

    where T(1)T(n) are the order statistics of T1,,Tn, and (βTZ)(1)(βTZ)(n) are the order statistics of βTZ1,,βTZn. Here, (βTZ)(i) is also known as the (i/n)th sample quantile of βTZ. However, T(i) is not fully observed due to right censoring, and naively treating Y(i), the order statistic of the observed time Y, as T(i) would cause bias.

    Denote Xβ=G(βTZ). Let FXβ(t)=Pr{G(βTZ)t} be the cumulative distribution function of the survival time Xβ. Let QT(τ)=inf{y:FT(y)τ} be the τ-quantile of FT(), and QXβ(τ)=inf{y:FXβ(y)τ} be the τ-quantile of FXβ. Motivated by [14], we define the objective function Mn(β)

    Mn(β)=Knk=1δk{ˆQT(τk)ˆQXβ(τk)}2I(αLτkαU), (2.2)

    where αLτ1<<τKnαU<1 are Kn quantile points, 0<δk=τkτk1, Knn, and Kn, max{δk}0 as n, ˆQT(τ) is the estimated τ-quantile with right-censored observations, and ˆQXβ(τ) is the sample τ-quantile of G(βTZ1),,G(βTZn). Here we confine the range of study in [0,τU], where τU(0,1) is a deterministic constant subject to certain identifiability constraints due to censoring. By matching censored quantiles, the estimator defined in Eq (2.2) forces the distribution FXβ to be as close as possible to the target distribution FT. Define ˆβ as a minimizer of minβMn(β). We call the proposed estimator ˆβ as the matching censored quantiles (MCQ) estimator.

    Remark 1. The proposed MCQ method has certain similarity with the idea of maximum rank correlation (MRC) estimator [17,18] which is given by minimizing

    ijI(Ti>Tj)I(βTZi>βTZj).

    The MRC approach also matches the orders of event times and covariate effects. However, there are essential differences between MCR and MCQ. The MRC method aims to match only the order of event times and covariate effects, not the quantiles, leading to a clear difference with MCQ in the form of objective functions. The objective function of MRC is a U-statistics, while the objective function of MCQ is a simple square summation. The MCQ method focus on minimizing the distance of the quantiles, so it allows the occurence of mismatch at some orders while the MRC method does not allow any mismatch. When there exist missing observations in Z, the MCQ method works normally but the MRC fails. Khan and Tamer [19] proposed a partial rank estimation (PRE) procedure which was a generalization of [17,18] for censored data. In Section 4, we compare the performance of the proposed method with that of the PRE method.

    The key to construct Eq (2.2) is to estimate the quantiles {QT(τk):k=1,,Kn}, for which the redistribution-of-mass technique (e.g., [8,10,20]) is adopted. This method redistributes the mass of each censored observation to Y+, where Y+ is a sufficiently large constant. We start with constructing an augmented data set {(Yi,Δi,Zi),i=1,,n+nc}, where {(Yi,Δi,Zi),i=1,,n} represent the original data, and {(Yi=Y+,Δi=0,Zi),i=n+1,,n+nc} are nc pseudo paired observations corresponding to the censored data.

    For the case of conditional independent censoring, given a fixed quantile τ, we define the local weight function as

    wi(FT;Zi,τ)={1, if Δi=1 or FT(Yi|Zi)>τ,τFT(Yi|Zi)1FT(Yi|Zi), if Δi=0 and FT(Yi|Zi)<τ, (2.3)

    for i=1,,n. Here, FT(t|Z) is the cumulative distribution function of T given Z. Let {w1(FT;Zi,τ),,wn(FT;Zi,τ),1wc1(FT;Zi,τ),,1wcnc(FT;Zi,τ)} be the weights assigned to the augmented data, where {c1,,cnc} are subscripts of the nc censored observations. Using these weights, we can estimate QT(τ) by

    ˆQT(τ)=inf{t:1nni=1[wi(FT;Zi,τ)I(Yit)+{1wi(FT;Zi,τ)}I(Yt)]τ}. (2.4)

    In practice, FT(t|Z) is unknown, and thus need to be estimated. Using the method in [21], we can estimate FT(t|Z) nonparametrically by ˆFT(t|Z=z)=1ˆST(t|Z=z) with ˆST(t|Z=z) being the local Kaplan-Meier estimator,

    ˆST(t|Z=z)=ni=1{1Bni(z)nk=1I(YkYi)Bnk(z)}I(Yit,Δi=1), (2.5)

    where Bni(z)=Kp{(zZi)/hn}/ni=1Kp{(zZi)/hn} is the Nadaraya-Watson type of weight, Kp(zi)=pj=1K(zij), K() is a univariate density kernel function, and hn is the bandwidth that converges to zero as n.

    For the case of independent censoring, we can still use the above framework for conditional independent censoring, and we only need to change the Bni(z) in Eq (2.5) as Bni(z)=1/n, for all i. In this case, ˆST(t|Z=z)=ˆST(t) exactly reduces to the Kaplan-Meier estimator, and the τ-quantile estimator by Eq (2.4) is equivalent to ˆQKM(τ)=inf{y:ˆFKM(y)τ}, where ˆFKM equals to 1 minus the Kaplan-Meier estimator.

    Since H() is a known and strictly monotonic function, in practical computation, people commonly assume H is from the class of Box-Cox transformation functions with a parameter λ as follows

    Hλ(t)={tλ1λ,λ>0,log(t),λ=0, (2.6)

    or other class of transformation functions such as logarithmic transformation function (Cheng et al. [3]). If there is no specific claim in the sequel, we assume Hλ=G1λ is from the Box-Cox transformations class in default.

    Let Xβ,λ=Gλ(βTZ), then, correspondingly, Eq (2.2) can be rewritten as

    Mn(β|λ)=Knk=1δk{ˆQT(τk)ˆQXβ,λ(τk)}2I(αLτkαU), (2.7)

    where QXβ,λ(τ)=inf{y:FXβ,λ(y)τ} and FXβ,λ(t)=Pr{Gλ(βTZ)t}. Let U() be the probability distribution function of the random variable FT(Xβ,λ). If T and Xβ,λ have the same distribution, FT(Xβ,λ) is a random variable uniformly distributed on the interval [0,1], hence U(x)=x, for x[0,1]. We define a measurement for the goodness of match as

    ρ=11210|dU(x)dx|. (2.8)

    It is obvious that ρ[0,1], and ρ=1 if and only if the matching is perfect in the sense that T and Xβ,λ have exactly the same distribution. When the difference between dU(x) and 1 increases, ρ decreases. Hence the larger the difference between the distributions of T and Xβ,λ, the smaller the value of ρ.

    Let ˆFT(t)=n1ni=1I(Tit) if there is no censoring, otherwise ˆFT(t)=ˆFKM(t), where ˆFKM(t)=1ˆSKM(t) and ˆSKM is the Kaplan-Meier estimator. Denote Vi=ˆFT(Xˆβ,ˆλ), and define

    ˆρ(ˆβ;ˆλ,k)=112n/ks=1|Dsk/n|,1kn, (2.9)

    where Ds=n1ni=1I{(s1)k/n<Visk/n}, and n/k represents the largest integer smaller than or equal to n/k.

    Considering that ˆρ can be used to measure the goodness of match between the distribution of Xˆβ,ˆλ and that of T, we shall use ˆρ as a criterion to choose the optimal value of λ for the transformation link functions in the sequel. We present the details of the proposed algorithm are as follows.

    Step 1. Given λ(1)=0, we update β by

    ˆβ(1)=argminβMn(β|λ(1))

    using the coordinate descent algorithm.

    Step 2. Calculate the value of goodness measurement of match, ˆρ(1), based on λ(1) and the obtained ˆβ(1).

    Step 3. Repeat Step 1 and Step 2 rest on the λ grid-searched in [0,L] with 0.1 as footstep, where L is a positive constant. At the same time, record all the values of {(λ(m),ˆβ(m),ˆρ(m)):m=1,,}, where stands for the number of the grid points of λ in [0,L].

    Step 4. Finally, take (λ(m),ˆβ(m)) with m corresponding to the maximum ˆρ(m) among all as the estimate (ˆλ,ˆβ).

    With the estimators (ˆλ,ˆβ), we then estimate the survival probability of T using

    ˆSXβ,λ(t)=11nni=1I{Gˆλ(ˆβTZi)t}.

    The computation in Step 1 involves bandwidth selection, which is critical for the local Kaplan-Meier estimator. In our numerical study, we use the leave-q-out cross-validation method on quantiles to choose hn. Specifically, we take Knq quantiles as the training set and the remaining q quantiles as the validation set (denoted as Vq). Given λ, we minimize Eq (2.7) using the Knq training quantiles, and then use the resulting coefficients ˆβTraining to predict the matching error at the validation quantiles by calculating the loss,

    τkVqδk{ˆQT(τk)ˆQXˆβTraining|λ(τk)}2I(αLτkαU).

    Repeat the above procedure and calculate the averaged prediction error until all the quantiles are scanned through. The bandwidth hn that yields the smallest averaged prediction error is selected. We set q=1 for the sequel numerical examples.

    Given the value of covariate Znew of a new patient and the obtained estimates ˆβ and ˆλ, we can predict the disease risk of a patient using the proposed method with following procedure:

    (ⅰ) Calculate the value of ˆβTZnew and denote it as tnew, calculate the empirical quantile τnew of tnew among the values of {ˆβTZi:i=1,,n}.

    (ⅱ) Calculate ˆS(t)=ˆP(T>t)=1n1ni=1I{Gˆλ(ˆβTZi)t}.

    Then, we predict the disease time for the patient by the value of ˆQT(τnew) and the disease risk by ˆS(tnew). In the sequel, we mainly interested in predicting the disease risk.

    In a general setting, suppose we are interested in matching a part of the target distribution, such as the segment between the αLth quantile and the αUth quantile, where αL and αU are prefixed and satisfy 0αL<αU<1. Let M(β)=αUαL{QT(τ)QXβ(τ)}2dτ. Define β0=argminβM(β), then β0 can be regarded as the theoretical true value to be estimated, although β0 may not be unique. Similar to the theoretical counterpart β0, the estimator ˆβ may not be unique either. We show below that Mn(β) converges to M(β) which is equivalent to show that the distribution of Xˆβ provides an optimal approximation to the distribution of T. Denote B={β:M(β)=M(β0)}, where is the Euclidean norm.

    Denote FC(t)=P(Ct) as the cumulative distribution function of censoring time C. Denote fξ() and fξ() as the density function and its first derivative function of a random variable ξ conditional on Z, respectively, where ξ could be T, C or Z. We impose the following regularity conditions.

    (C1) Assume B is a compact subsets of Rp+1, and T has a bounded support.

    (C2) The density functions fT(), fC(), fXβ(), fT(|Z) and fC(|Z) are uniformly bounded away from 0 and infinity, and FT(t) and FC(t) have uniformly bounded second-order partial derivatives with respect to Z.

    (C3) For any fixed β, it holds that

    supαLααU|fT{QT(α)}|<,infαLααUfT{QT(α)}>0, and supαLααU|fXβ{QXβ(α)}|<,infαLααUfXβ{QXβ(α)}>0.

    (C4) The kernel function K()0 has a compact support, K() is Lipschitz continuous of order 1 and satisfies K(u)du=1, uK(u)du=0, K2(u)du<, and |u|2K(u)du<.

    (C5) The bandwidth satisfies hn=O(nv), where 0<v<1/p.

    (C6)G is a thrice continuously differentiable and strictly increasing function.

    The first part of condition (C1) imposes a regular assumption on the true parameter space. Considering that the follow-up study is typically restricted to some limited time, the second part of condition (C1) which assumes T has a bounded support is also reasonable. Condition (C2) is necessary for the Kaplan-Meier estimator, and it shall be used to derive the consistency of the proposed estimators. Condition (C3) is the Kiefer condition [22] that ensures the uniform Bahadur–Kiefer bounds for empirical quantile processes with independent and identically distributed samples. Conditions (C4) and (C5) are regular assumptions for kernel-based smoothing estimators in terms of the bandwidth and the kernel function. Condition (C6) is satisfied by G(x)=(λx+1)1/λ1 with λ>0, which corresponds to H(x) being the Box-Cox transformation function. Under the conditions above, we have the following two theorems.

    Theorem 1. Under conditions (C1)–(C6), Mn(ˆβ)M(β0) in probability, and d(ˆβ,B)0 in probability.

    The consistency shown in Theorem 1 indicates that the distribution of G(ˆβTZ) shall provides an optimal approximation to the distribution of T when the sample size is sufficiently large, although both of them may not converge exactly to the true distribution FT. Proof of Theorem 1 is sketched in the Supplementary.

    To illustrate the finite sample performance of the proposed methods, we conduct the following two simulation studies.

    Example 1. Consider a normal error transformation model

    Hλ(Ti)=βTZi+ϵi,

    where Hλ() is a Box-Cox transformation function with parameter λ=0,0.5 or 1, Zi=(Zi1,Zi2)T, β=(2,1)T, Zi1 and Zi2 are independent and follow the normal distribution with mean 5 and standard deviation 1, and ϵi independently follows the standard normal distribution. The right censoring time Ci is generated independently from uniform distributions to yield the censoring rates of 20 and 40%, correspondingly.

    Example 2. We compare the proposed method with a regression method using Cox proportional hazard model under the samples generated from three different models:

     Model I:log(Ti)=βTZi+ϵi, Model II, III: Λ(t|Zi)=Hλ{Λ0(t)exp(βTZi)} with λ=0and1,respectively,

    where Hλ() is a logarithmic transformation function,

    Hλ(t)={log(1+λt)λ,λ>0,t,λ=0,

    Λ0(t)=t, Zi=(Zi1,Zi2)T, β=(2,1)T, Zi1, Zi2 and ϵi are independent and follow the standard normal distribution. The right censoring time Ci is generated independently from exponential distributions to yield the censoring rates of 20 and 40%, correspondingly. Specifically, Model Ⅰ corresponds to the log-normal AFT model, Model Ⅱ with λ=0 corresponds to the Cox's proportional hazards model, and Model Ⅲ corresponds to the proportional odds model.

    Example 3. Consider the same model and parameter settings as in Example 1 except that the censoring time Ci is generated independently from exponential distributions to yield the censoring rates of 20 and 40%, respectively. Besides, we let the values of Zis be missing completely at random with a missing rates of 20 and 40%, respectively, thus the simulated observations in Example 3 are unpaired.

    Example 4. To assess the robustness of the proposed method, we compare the proposed method with three alternative methods in this example. For convenience, we consider a same model log(Ti)=βTZi+ϵi, i=1,,n, used in Example 2. Consider two scenarios with ϵiN(0,1) and ϵit(3), respectively. We compare the proposed method with the PRE method, rank-based estimation (AFT.rank) [23], and the regular least squares estimation of AFT for censored data [24]. The last one is a non-robust method, so we choose it as a benchmark. To evaluate the accuracy of prediction of the proposed method, we independently generate 20%n testing samples from the model, and then predict the potential risk at each testing observation using the well established predictor using training data.

    In the simulation, we choose a quadratic (biweight) kernel function, K(x)=15/16(1x2)2I(|x|1), for the MCQ method. Other kernel functions, such as the Epanechnikov kernel, can also be used, while our experience shows that the numerical results by different kernel functions have little difference. We adopt a product kernel for the multivariate scenario. For example, we use K(x1,x2)=K(x1)K(x2) for bivariate cases, where K() is a univariate quadratic kernel function. We set αL=0 and αU=0.90. We take L=2 for searching the optimal transformation function. The simulation results are based on 1000 replications with sample sizes of 200,400 and 800.

    Table 1 presents the simulation results of Example 1, from which we can observe that the proposed MCQ method performs well overall on assessing the target distribution in terms of the estimated values at the prefixed τ-quantiles. Although the bias of the estimated survival rates tends to be increasing as the censoring rate increases under small sample sizes, the estimation accuracy is improved significantly as the sample size increases. On the other hand, from the estimated values of λ, the method used for searching the parameter λ in the function G based on the criterion ˆρ also works considerably well. Figure 1 shows the curves of the estimated ρ along with λ with the true values of λ equal to 0, 0.5 and 1.0, respectively. The plots show the clear modes of the maximum points around the true values.

    Table 1.  The empirical median of the estimated values ^Pr{Xˆβ,ˆλQT(τ)} and the estimated λ in Example 1 with different sample sizes and censoring rates (the values in the parentheses are standard deviations).
    Censoring rate = 0% Censoring rate = 20% Censoring rate = 40%
    True n=400 n=800 n=400 n=800 n=400 n=800
    τ=0.25 0.252(0.022) 0.250(0.015) 0.248(0.022) 0.250(0.017) 0.235(0.030) 0.240(0.026)
    τ=0.50 0.500(0.026) 0.502(0.019) 0.505(0.030) 0.506(0.020) 0.495(0.041) 0.496(0.033)
    τ=0.75 0.750(0.022) 0.750(0.015) 0.760(0.026) 0.758(0.020) 0.780(0.037) 0.765(0.030)
    λ=0.00 0.007(0.037) 0.000(0.006) 0.029(0.071) 0.006(0.035) 0.067(0.108) 0.034(0.095)
    τ=0.25 0.252(0.023) 0.250(0.016) 0.250(0.028) 0.251(0.022) 0.248(0.040) 0.249(0.029)
    τ=0.50 0.502(0.029) 0.502(0.021) 0.502(0.038) 0.502(0.029) 0.512(0.058) 0.509(0.042)
    τ=0.75 0.750(0.022) 0.750(0.015) 0.758(0.031) 0.754(0.022) 0.780(0.049) 0.770(0.034)
    λ=0.50 0.600(0.364) 0.495(0.259) 0.600(0.446) 0.520(0.352) 0.600(0.638) 0.550(0.485)
    τ=0.25 0.252(0.023) 0.250(0.016) 0.252(0.025) 0.251(0.017) 0.252(0.026) 0.252(0.019)
    τ=0.50 0.502(0.028) 0.502(0.021) 0.500(0.030) 0.502(0.023) 0.500(0.033) 0.501(0.025)
    τ=0.75 0.750(0.021) 0.750(0.015) 0.750(0.023) 0.750(0.017) 0.751(0.026) 0.751(0.018)
    λ=1.00 1.000(0.629) 1.000(0.471) 1.150(0.679) 1.000(0.507) 1.200(0.712) 1.000(0.543)

     | Show Table
    DownLoad: CSV
    Figure 1.  Curves of the estimated ρ along with λ under Example 1 with n=800 and the true values of λ equal to 0, 0.5 and 1.0, respectively.

    In Example 2, we compare the proposed method with the other two methods of estimating survival rates under three models. The results are summarized in Table 2. To be specific, 'Pro' is the prediction by the proposed matching censored quantiles method; 'Cox' stands of the survival predictor based on the maximum partial likelihood estimation method for proportional hazards model; 'K-M' is survival prediction using the Kaplan-Meier estimator. The K-M values can be regarded as the benchmark estimator, which are considerably accurate under all the scenarios. Naturally, the Cox method performs well under the true model while show underperforms with misspecified models. Compared to the Cox method, the proposed method performs relatively stable. In Example 3, we demonstrate that the proposed method can handle the unpaired observations caused by missing data. The simulation results in Table 3 show that MCQ performs well in assessing the values of FT at the 0.25-quantile, 0.50-quantile and 0.75-quantile, which demonstrates their special advantages in dealing with unpaired observations over the traditional methods.

    Table 2.  The empirical median of the estimated values ^Pr{TQT(τ)} with different models in Example 2 (the values in the parentheses are standard deviations).
    Censoring rate = 0% Censoring rate = 20% Censoring rate = 40%
    model τ n=400 n=800 n=400 n=800 n=400 n=800
    0.25 Pro. 0.252(0.022) 0.250(0.015) 0.248(0.022) 0.250(0.017) 0.235(0.030) 0.240(0.026)
    Cox 0.284(0.031) 0.281(0.021) 0.281(0.032) 0.278(0.022) 0.273(0.032) 0.274(0.022)
    K-M 0.252(0.021) 0.251(0.016) 0.254(0.021) 0.252(0.016) 0.252(0.021) 0.253(0.017)
    0.50 Pro. 0.500(0.026) 0.502(0.019) 0.505(0.030) 0.506(0.020) 0.495(0.041) 0.496(0.033)
    Cox 0.539(0.037) 0.541(0.026) 0.525(0.037) 0.527(0.026) 0.514(0.040) 0.516(0.028)
    K-M 0.501(0.028) 0.501(0.021) 0.503(0.029) 0.501(0.022) 0.500(0.029) 0.501(0.023)
    0.75 Pro. 0.750(0.022) 0.750(0.015) 0.760(0.026) 0.756(0.020) 0.780(0.037) 0.765(0.030)
    Cox 0.771(0.029) 0.769(0.022) 0.752(0.031) 0.751(0.023) 0.739(0.039) 0.736(0.029)
    K-M 0.750(0.023) 0.750(0.016) 0.751(0.025) 0.751(0.019) 0.751(0.030) 0.753(0.023)
    0.25 Pro. 0.260(0.023) 0.263(0.015) 0.259(0.024) 0.262(0.017) 0.260(0.030) 0.264(0.022)
    Cox 0.248(0.033) 0.246(0.020) 0.248(0.033) 0.246(0.020) 0.249(0.034) 0.247(0.020)
    K-M 0.250(0.023) 0.249(0.014) 0.251(0.023) 0.249(0.015) 0.250(0.023) 0.250(0.014)
    0.50 Pro. 0.510(0.025) 0.514(0.016) 0.525(0.029) 0.519(0.018) 0.524(0.046) 0.515(0.032)
    Cox 0.506(0.040) 0.502(0.028) 0.504(0.040) 0.501(0.029) 0.497(0.041) 0.500(0.029)
    K-M 0.501(0.026) 0.499(0.019) 0.501(0.026) 0.499(0.020) 0.501(0.028) 0.501(0.021)
    0.75 Pro. 0.750(0.022) 0.751(0.013) 0.764(0.032) 0.757(0.024) 0.765(0.042) 0.756(0.031)
    Cox 0.746(0.034) 0.750(0.025) 0.749(0.035) 0.748(0.025) 0.744(0.037) 0.747(0.027)
    K-M 0.747(0.022) 0.748(0.017) 0.750(0.023) 0.749(0.018) 0.751(0.028) 0.745(0.022)
    0.25 Pro. 0.260(0.022) 0.259(0.015) 0.255(0.023) 0.258(0.015) 0.255(0.028) 0.259(0.018)
    Cox 0.260(0.032) 0.263(0.019) 0.258(0.033) 0.260(0.020) 0.255(0.033) 0.258(0.020)
    K-M 0.249(0.023) 0.250(0.015) 0.249(0.023) 0.251(0.015) 0.246(0.023) 0.251(0.015)
    0.50 Pro. 0.505(0.024) 0.503(0.017) 0.512(0.028) 0.510(0.019) 0.519(0.050) 0.509(0.036)
    Cox 0.526(0.036) 0.526(0.023) 0.521(0.036) 0.517(0.024) 0.513(0.037) 0.509(0.025)
    K-M 0.506(0.025) 0.499(0.018) 0.504(0.025) 0.501(0.018) 0.505(0.027) 0.502(0.019)
    0.75 Pro. 0.745(0.025) 0.742(0.015) 0.755(0.034) 0.751(0.025) 0.760(0.039) 0.757(0.029)
    Cox 0.764(0.029) 0.764(0.019) 0.751(0.031) 0.752(0.021) 0.736(0.039) 0.740(0.027)
    K-M 0.752(0.022) 0.748(0.015) 0.750(0.023) 0.748(0.015) 0.748(0.032) 0.749(0.022)
    NOTE: 'Pro' indicates the proposed method; 'Cox' indicates the proportional hazards model; 'K-M' indicates the Kaplan-Meier estimator. The τ-quantiles (QT(0.25),QT(0.5),QT(0.75)) for prediction in model Ⅰ, Ⅱ and Ⅲ are (0.2594,0.9993827,3.8538), (0.1411,0.6041,2.4361) and (0.1930,0.9995,5.1773), respectively.

     | Show Table
    DownLoad: CSV
    Table 3.  The empirical mean of the estimated values of ^Pr{Xˆβ,ˆλQT(τ)} by the proposed method under Example 3 with unpaired observations. (Median absolute deviation values are given in the parentheses).
    n=400 n=800
    miss. c% τ λ=0 λ=1 λ=0 λ=1
    0% 0 0.25 0.250(0.022) 0.250(0.022) 0.251(0.017) 0.250(0.015)
    0.50 0.500(0.024) 0.502(0.026) 0.501(0.019) 0.500(0.019)
    0.75 0.750(0.022) 0.750(0.022) 0.751(0.015) 0.750(0.015)
    20 0.25 0.250(0.022) 0.250(0.022) 0.251(0.015) 0.250(0.017)
    0.50 0.507(0.026) 0.500(0.022) 0.501(0.019) 0.505(0.020)
    0.75 0.760(0.026) 0.750(0.022) 0.751(0.017) 0.756(0.019)
    40 0.25 0.250(0.026) 0.250(0.022) 0.251(0.015) 0.250(0.019)
    0.50 0.510(0.033) 0.499(0.024) 0.501(0.019) 0.506(0.022)
    0.75 0.760(0.033) 0.748(0.026) 0.750(0.019) 0.756(0.022)
    20% 0 0.25 0.250(0.023) 0.250(0.023) 0.252(0.014) 0.252(0.016)
    0.50 0.500(0.028) 0.500(0.023) 0.500(0.019) 0.500(0.019)
    0.75 0.750(0.023) 0.750(0.023) 0.752(0.016) 0.752(0.016)
    20 0.25 0.253(0.023) 0.250(0.023) 0.252(0.016) 0.252(0.016)
    0.50 0.503(0.023) 0.500(0.023) 0.502(0.019) 0.503(0.019)
    0.75 0.753(0.023) 0.750(0.023) 0.752(0.016) 0.752(0.016)
    40 0.25 0.250(0.023) 0.250(0.023) 0.252(0.014) 0.250(0.019)
    0.50 0.509(0.032) 0.500(0.028) 0.502(0.019) 0.506(0.021)
    0.75 0.759(0.032) 0.747(0.028) 0.750(0.019) 0.756(0.023)

     | Show Table
    DownLoad: CSV

    In Example 4, we compare the robustness and efficiency of the proposed method with three alternative approaches in terms of estimation and prediction. The simulation results are summarized in Table 4. From the estimation results, we see that the proposed method and the AFT.rank method perform almost equally well in terms of robustness and better than the other two methods under the heavy-tail scenario of t(3) model errors. Khan's method provides larger bias estimation than the proposed method, which shows that the proposed method could gain more flexibility by allowing certain mismatch of orders at some quantiles and hence can improve both the estimation and prediction of the disease risk. Moreover, we also see that the proposed method has a clear advantage in efficiency compared to the traditional robust method based on rank estimation. From the prediction results in Table 4, we know the prediction accuracy of proposed method is better than that of Khan's method.

    Table 4.  The empirical median of the estimated values ^Pr{TQT(τ)} and the empirical bias of prediction of disease risk with different approaches in Example 4 (the values in the parentheses are standard deviations).
    Estimation
    Censoring rate = 0% Censoring rate = 20%
    ϵ n τ% Prop Khan AFT.rank AFT Prop Khan AFT.rank AFT
    N(0,1) 200 25 25.3(2.7) 20.0(3.1) 23.0(8.6) 25.1(2.5) 25.1(3.2) 19.8(3.3) 23.5(10.6) 22.4(3.3)
    50 49.1(4.0) 46.9(4.7) 50.6(11.8) 50.0(3.1) 49.9(4.1) 46.4(4.8) 50.7(14.5) 50.9(3.7)
    75 74.4(3.3) 75.0(4.2) 77.3(9.6) 75.1(2.6) 76.5(3.5) 74.5(4.6) 76.8(11.7) 78.7(3.3)
    400 25 25.1(2.1) 19.8(2.8) 22.0(6.2) 25.0(1.8) 24.1(2.5) 19.6(2.9) 22.1(6.7) 21.6(2.3)
    50 49.5(2.7) 46.4(4.3) 49.7(8.9) 49.9(2.1) 49.2(3.2) 46.3(4.5) 49.9(9.7) 49.8(2.7)
    75 74.6(2.4) 74.7(4.1) 77.3(7.3) 75.0(1.9) 75.6(2.6) 74.6(4.1) 77.2(8.1) 78.0(2.4)
    t(3) 200 25 24.8(3.0) 18.3(3.6) 21.3(10.1) 23.0(2.8) 26.4(3.3) 18.3(3.4) 21.4(11.1) 19.7(3.6)
    50 49.1(3.8) 46.8(5.3) 50.3(13.9) 50.3(4.0) 48.6(4.7) 47.2(5.1) 50.5(15.2) 50.2(4.4)
    75 75.2(3.3) 77.1(4.6) 78.8(10.6) 79.3(3.5) 72.9(3.6) 77.6(4.6) 78.4(12.0) 80.1(3.7)
    400 25 25.1(2.4) 18.2(3.0) 20.9(6.7) 23.7(2.4) 26.5(2.6) 18.1(2.9) 21.3(7.4) 19.7(2.5)
    50 49.7(2.9) 46.9(4.4) 51.1(9.9) 50.0(3.1) 49.1(3.3) 46.9(4.6) 51.8(11.1) 50.6(3.2)
    75 75.1(2.4) 77.3(4.0) 80.4(7.5) 80.5(2.7) 73.3(2.4) 77.2(4.3) 80.4(8.7) 80.9(2.9)
    Prediction
    Censoring rate = 0% Censoring rate = 20%
    ϵ n Bias Prop Khan AFT.rank AFT Prop Khan AFT.rank AFT
    N(0,1) 200 ˆS(tnew) 1.0(3.6) 4.6(6.1) 0.3(3.4) 0.3(3.4) 0.3(3.9) 5.5(6.2) 0.0(3.8) 0.0(3.8)
    400 ˆS(tnew) 0.1(2.5) 4.9(6.1) 0.2(2.4) 0.2(2.2) 0.1(2.8) 5.5(6.4) 0.2(2.7) 0.3(2.4)
    t(3) 200 ˆS(tnew) 0.4(3.7) 3.8(6.4) 0.2(3.6) 0.1(3.3) 0.9(3.1) 4.3(7.1) 0.1(3.3) 0.2(3.8)
    400 ˆS(tnew) 0.4(2.7) 5.8(6.6) 0.2(2.8) 0.2(2.4) 0.2(2.5) 5.6(5.7) 0.1(2.4) 0.1(2.8)
    NOTE: 'Prop' indicates the proposed method; 'Khan' indicates the PRE method with the idea of maximum rank correlation; 'AFT.rank' indicates the rank-based estimation; 'AFT' indicates the least squares estimation by Jin et al. [24].

     | Show Table
    DownLoad: CSV

    We apply the proposed methods to analyze Veterans' administration lung cancer data collected from the patients with advanced inoperable lung cancer [25]. This data set consists of 137 patients who were randomized to either a standard or test chemotherapy, among them 128 were followed to death. In this study, one primary endpoint for the therapy comparison is the time to death. In addition to the treatment indicator, several covariates are included, such as the Karnofsky performance score (Karnofsky), the time in months from the diagnosis to randomization (diagtime), the prior therapy (yes or no), and the patient's age in years and the lung cancer cell type (squamous cell = 1, small cell = 2, adenocarcinoma cell = 3, large cell = 4). According to existing literature, all the above covariates are important factors affecting the survival time.

    The goal is to estimate β and λ such that the distribution of Gλ(βTZ) matches the distribution of T, i.e., the survival time in days. The covariate Zi=(1,Zi1,,Zi8)T are defined as follows: Zi1=age/100, Zi2=Karnofsky/10, Zi3=diagtime/100, and Zi4=0 if no prior therapy and 1 otherwise; Zij=1, j=5,6,7, if the cell type is squamous, small or large, respectively, and 0 otherwise; Zi8=1 if the patient is given the test chemotherapy treatment, otherwise Zi8=0. To apply the proposed methods, we set αL=0 and αU=ˆFKM(1000)=0.985 for the MCQ method. Considering the fact that the estimated coefficients by the proposed method may not be unique, we compare the estimated survival curve by the proposed methods with that by the Kaplan-Meier estimator and the Cox proportional hazards model. To be specific, we calculate ˆSˆβ(t)=1n1ni=1I{Gˆλ(ˆβTZi)t} using the estimated regression coefficients ˆβ. Then, we use ˆSˆβ to assess or approximate the survival probability of T. For the Cox proportional hazards model, we use ˉS(t|Z)=n1ni=1exp{ˆΛ0(t)exp(ˆβTCoxZi)} to estimate the survival function of T, where ˆΛ0 is the estimated baseline cumulative hazard function, and ˆβCox is the estimated regression coefficient.

    The estimated value of λ for the transformation function Gλ is 0.4 which corresponds to the value of ρ=0.892. The estimated values of ρ indicate that the proposed method performs considerably well in matching ˆFKM. Table 5 presents the estimated survival rates of the survival time with 95% confidence intervals (CI), and Figure 3 displays the estimated survival curves of T by MCQ with 95% pointwise confidence bands (CB). Here, both the 95% CIs and the 95% CBs are constructed by the 0.025 and 0.975 quantiles of the estimated survival rates through the bootstrap method with 1000 bootstrap samples. Overall, the estimated survival curves by MCQ and KM are considerably close to each other, except for minor differences at the tail. Compared with the survival curve by the Cox proportional hazards model, the MCQ curve is much closer to the Kaplan-Meier curve, which indicates that the estimated survival probability by the proposed methods might be more accurate than that by the Cox proportional hazards model.

    Table 5.  Estimated survival rates of the survival time with 95% confidence intervals (CI) for Veterans' administration lung cancer data analysis.
    T (in days) Est. 95% CI
    50 0.617 (0.540,0.706)
    100 0.451 (0.350,0.519)
    200 0.226 (0.153,0.294)
    300 0.120 (0.080,0.206)
    400 0.090 (0.015,0.139)
    500 0.053 (0.000,0.097)
    600 0.015 (0.000,0.067)

     | Show Table
    DownLoad: CSV
    Figure 2.  Curve of the estimated values of ρ along with λ.
    Figure 3.  The estimated curves of survival rates by different methods.

    Remark 2. Here we assume, without any proof, that the proposed estimator converges to an asymptotic distribution, hence we could use the bootstrap method under this assumption. In other words, the bootstrap method used here is not rigorous from the theoretical point of view.

    In this paper, we propose a matching censored quantiles estimator to study the relationship between the observed event times and the covariates in the presence of right censoring. The proposed method provides a new option to predict the disease risks for survival events of interest.

    Under the MCQ framework, we adopt a locally weighted approach to estimate the censored quantiles. Other options such as the inverse probability weighting or the Buckley-James method [1] can also be considered. Yet, it still lacks efficient approaches for statistical inference and hypothesis testing on the obtained estimators which warrant further research. Last but not least, new algorithms are also needed to combine the proposed methods with sparse model selection using penalty functions.

    This research was supported by the Fundamental Research Funds for the Central Universities, Beijing Natural Science Foundation (No. 1204031), and the National Natural Science Foundation of China (No. 11901013).

    All authors declare no conflicts of interest in this paper.

    Lemma 1. Under conditions (C2), (C4) and (C5), we have

    suptsupZ|ˆFT(t|Z)FT(t|Z)|=op(1),

    where 0<l0<1/4.

    Proof. Lemma 1 follows from the result of Theorem 2.1 of [26] and Lemma A.1 of [27].

    Lemma 2. If Mn(ˆβn)M(β0) in probability, where the estimator ˆβn is defined by minimizing Mn(β), then d(ˆβn,B)0 in probability as n.

    Proof. To prove this lemma is equivalent to prove that Pr{d(ˆβn,B)ϵ}0 for any constant ϵ>0. Suppose there exists an ϵ>0 such that limsupnPr{d(ˆβn,B)ϵ}>0, then there exists a subsequence of {ˆβn}, denoted by {ˆβnk}, such that limkPr{d(ˆβnk,B)ϵ}=η>0. Define Bϵ={β:d(β,B)ϵ}, hence Bϵ is a compact set. Because B={β:M(β)=M(β0)}, then infβBϵM(β)=η+M(β0). Hence, it holds

    limkPr{|Mnk(ˆβnk)M(ˆβnk)|<η/2}=1.

    Hence, we have

    Mnk(ˆβnk)M(ˆβnk)η/2infβBϵM(β)η/2>M(β0)+η/2>M(β0).

    This contradicts to the assumption that Mn(ˆβn)M(β0) in probability. This completes the proof of Lemma 2.

    Proof of Theorem 1. By the definition of Mn(β), we have

    |Mn(β)M(β)|2Knk=1δk{ˆQT(τk)QT(τk)}2I(αLτkαU)+C1Knk=1δk|ˆQT(τk)QT(τk)|I(αLτkαU)+2Knk=1δk{ˆQXβ(τk)QXβ(τk)}2I(αLτkαU)+C2Knk=1δk|ˆQXβ(τk)QXβ(τk)|I(αLτkαU)+|Knk=1δk{QT(τk)QXβ(τk)}2I(αLτkαU)αUαL{QT(τ)QXβ(τ)}2dτ|,

    where C1>0 and C2>0 are some constants. According to the definition of the Riemann integral, the last term on the right-hand side of the above equation tends to 0 as Kn and max{δk}0 for any finite β. We only need to consider the rear terms.

    Firstly, because Knk=1δk{ˆQT(τk)QT(τk)}2I(αLτkαU)=αUαL{ˆQT(t)QT(t)}2dt+o(1), and

    αUαL{ˆQT(t)QT(t)}2dt=αUαL(QT[FT{ˆQT(t)}]QT(t))2dt=αUαL{dQT(t)dt}2[FT{ˆQT(t)}t]2dt,

    where the last equation is by the mean value theorem, dQT(t)/dt is the first derivative of QT(t), and t is a point between FT{ˆQT(t)} and t. By conditions (C1)–(C3), we know |dQT(t)/dt| can be bounded by a positive constant. According to Lemma 1, the proof of Theorem 1 in [10] and the dominated convergence theorem, we have Knk=1δk{ˆQT(τk)QT(τk)}2I(αLτkαU)=oP(1). Using the same argument, we can also conclude that Knk=1δk|ˆQT(τk)QT(τk)|I(αLτkαU)=oP(1). Second, note that the condition (C3) entails that Xβ has a bounded support for any β, even in the extreme case with αL=0 and αU close to 1. Using the same argument above and the Glivenko-Cantelli theorem, for any fixed β we can also obtain Knk=1δk{ˆQXβ(τk)QXβ(τk)}2I(αLτkαU)=oP(1) and Knk=1δk|ˆQXβ(τk)QXβ(τk)|I(αLτkαU)=oP(1). Hence, |Mn(β)M(β)|0 in probability as n for any fixed and finite β.

    By conditions (C1) and (C2), the matching censored quantiles estimator defined by Eq (2.2) is finite in Rp+1, hence there exists a compact neighborhood BRp+1 such that ˆβB. Next, we show supβB|Mn(β)M(β)|0 in probability. Because M(β) is a continuous function with respect to β. According to the Heine–Borel theorem, for any ϵ>0, there exist finite elements β1,,βmB, m is a finite integer, such that ββj<Cϵ and |M(β)M(βj)|<ϵ, where C is a constant, j=1,,m. By conditions (C2), (C3) and the Glivenko-Cantelli theorem, we have

    |Mn(β)M(β)||Mn(β)Mn(βj)|+|Mn(βj)M(βj)|+|M(βj)M(β)|O(ϵ)+|Mn(βj)M(βj)|.

    Hence, supβB|Mn(β)M(β)|O(ϵ)+mj=1|Mn(βj)M(βj)|, by the arbitrariness of ϵ, we conclude that supβB|Mn(β)M(β)|0 in probability for a compact neighborhood BRp+1. Finally, by the inequation

    |Mn(ˆβ)M(ˆβ)||Mn(ˆβ)M(β0)||Mn(β0)M(β0)|,

    and the fact that |Mn(β0)M(β0)|0 and |Mn(ˆβ)M(ˆβ)|0 in probability, it holds that |Mn(ˆβ)M(β0)|0 in probability. Using the same argument in the proof of Lemma 2, the second part of Theorem 1 is proved.



    [1] The Center for Systems Science and Engineering (CSSE) at JHU. Available from: https: //systems.jhu.edu/ (accessed on 15 April 2020).
    [2] National Health Commission of the People's Republic of China (in Chinese), 2020. Available from: http://www.nhc.gov.cn/xcs/yqtb/202003/097e6e91ecb6464ea69fd1a324c9b 1b4.shtml.
    [3] D. Ni, A comparative study of the two strategies for COVID-19, 2020. Available from: http: //cn.chinadaily.com.cn/a/202003/19/WS5e731e02a3107bb6b57a7913.html.
    [4] Korea Centers for Disease Control and Prevention, 2020. Available from: https://www.cdc.go.kr/board/board.es?mid=a30402000000&bid=0030.
    [5] Ministry of Health, Labour and Welfare, 2020. Available from: https://www.mhlw.go.jp/stf/seisakunitsuite/bunya/0000121431_00086.html.
    [6] Spanish Ministry of Health, 2020. Available from: https://www.isciii.es/QueHacemos/Servicios/VigilanciaSaludPublicaRENAVE/EnfermedadesTransmisibles/Paginas/ InformesCOVID-19.aspx.
    [7] World Health Organization, 2020. Available from: https://www.who.int/emergencies/diseases/novel-coronavirus-2019/situation-reports.
    [8] B. Xu, B. Gutierrez, S. Mekaru, K. Sewalk, L. Goodwin, A. Loskill, et al., Epidemiological data from the COVID-19 outbreak, real-time case information, Sci. Data, 7 (2020), 1-6.
    [9] E. Goldstein, J. Dushoff, J. Ma, J. B. Plotkin, D. D. Earn, M. Lipsitch, Reconstructing influenza incidence by deconvolution of daily mortality time series, Proc. Natl. Acad. Sci., 106 (2009), 21825-21829.
    [10] Q. Li, X. Guan, P. Wu, X. Wang, L. Zhou, Y. Tong, et al., Early transmission dynamics in Wuhan, China, of novel coronavirus-infected pneumonia, N. Engl. J. Med., 382 (2020), 1199-1207.
    [11] H. Nishiura, M. G. Roberts, Estimation of the reproduction number for 2009 pandemic influenza A(H1N1) in the presence of imported cases, Euro Surveill., 15 (2010), 19622.
    [12] A. Cori, N. M. Ferguson, C. Fraser, S. Cauchemez, A new framework and software to estimate time-varying reproduction numbers during epidemics, Am. J. Epidemiol., 178 (2013), 1505-1512.
    [13] H. Nishiura, N. M. Linton, A. R. Akhmetzhanov, Serial interval of novel coronavirus (COVID-19) infections, Int. J. Infect. Dis., 93 (2020), 284-286.
    [14] T. Gnyani, C. Kremer, D. Chen, A. Torneri, C. Faes, J. Wallinga, et al., Estimating the generation interval for COVID-19 based on symptom onset data, medRxiv, (2020). doi: https://doi.org/10.1101/2020.03.05.20031815.
    [15] C. You, Y. Deng, W. Hu, J. Sun, Q. Lin, F. Zhou, et al., Estimation of the time-varying reproduction number of COVID-19 outbreak in China, Int. J. Hyg. Environ. Health, 228 (2020), 113555.
    [16] S. Zhao, Estimating the time interval between transmission generations when negative values occur in the serial interval data: using COVID-19 as an example, Math. Biosci. Eng., 17 (2020), 3512-3519.
    [17] K. Wang, S. Zhao, H. Li, Y. Song, L. Wang, M. H. Wang, et al., Real-time estimation of the reproduction number of the novel coronavirus disease (COVID-19) in China in 2020 based on incidence data, Ann. Transl. Med., 8 (2020), 689.
    [18] S. Flaxman, S. Mishra, A. Gandy, H. Unwin, H. Coupland, T. A. Mellan, et al., Report 13: Estimating the number of infections and the impact of nonpharmaceutical interventions on COVID-19 in 11 European countries. (2020). doi: https://doi.org/10.25561/77731.
    [19] Reuters. Available from: https://graphics.reuters.com/CHINA-HEALTH-SOUTHKOREA-CLUSTERS/0100B5G33SB/index.html (accessed on 15 April 2020).
    [20] S. Zhao, Q. Lin, J. Ran, S. S. Musa, G. Yang, W. Wang, et al., Preliminary estimation of the basic reproduction number of novel coronavirus (2019-nCoV) in China, from 2019 to 2020: A data-driven analysis in the early phase of the outbreak, Int. J. Infect. Dis., 92 (2020), 214-217.
    [21] S. Sanche, Y. T. Lin, C. Xu, E. Romero-Severson, N. Hengartner, R. Ke, High contagiousness and rapid spread of severe acute respiratory syndrome coronavirus 2, Emerg. Infect. Dis., 26 (2020), 1470-1477.
    [22] B. Tang, X. Wang, Q. Li, N. L. Bragazzi, S. Tang, Y. Xiao, et al., Estimation of the transmission risk of the 2019-nCoV and its implication for public health interventions, J. Clin. Med., 9 (2020), 462.
    [23] Y. Xiao, B. Tang, J. Wu, R. A. Cheke, S. Tang, Linking key intervention timing to rapid decline of the COVID-19 effective reproductive number to quantify lessons from mainland China, Int. J. Infect. Dis., 97 (2020), 296-298.
    [24] E. Shim, A. Tariq, W. Choi, Y. Lee, G. Chowell, Transmission potential and severity of COVID-19 in South Korea, Int. J. Infect. Dis., 93 (2020), 339-344.
    [25] J. Hwang, H. Park, J. Jung, S. Kim, N. Kim, Basic and effective reproduction numbers of COVID-19 cases in South Korea excluding Sincheonji cases, medRxiv, (2020).
    [26] S. Kim, Y. D. Jeong, J. H. Byun, G. Cho, A. Park, J. H. Jung, et al., Evaluation of COVID-19 epidemic outbreak caused by temporal contact-increase in South Korea, Int. J. Infect. Dis., 96 (2020), 454-457.
    [27] B. Tang, F. Xia, S. Tang, N. L. Bragazzi, Q. Li, X. Sun, et al., The effectiveness of quarantine and isolation determine the trend of the COVID-19 epidemics in the final phase of the current outbreak in China, Int. J. Infect. Dis., 95 (2020), 288-293.
    [28] H. Legido-Quigley, N. Asgari, Y. Teo, G. M. Leung, H. Oshitani, K. Fukuda, et al., Are high-performing health systems resilient against the COVID-19 epidemic?, The Lancet, 395 (2020), 848-850.
  • This article has been cited by:

    1. Nuno Canha, Catarina Teixeira, Mónica Figueira, Carolina Correia, How Is Indoor Air Quality during Sleep? A Review of Field Studies, 2021, 12, 2073-4433, 110, 10.3390/atmos12010110
    2. Emerson M. Wickwire, Scott G. Williams, Thomas Roth, Vincent F. Capaldi, Michael Jaffe, Margaret Moline, Gholam K. Motamedi, Gregory W. Morgan, Vincent Mysliwiec, Anne Germain, Renee M. Pazdan, Reuven Ferziger, Thomas J. Balkin, Margaret E. MacDonald, Thomas A. Macek, Michael R. Yochelson, Steven M. Scharf, Christopher J. Lettieri, Sleep, Sleep Disorders, and Mild Traumatic Brain Injury. What We Know and What We Need to Know: Findings from a National Working Group, 2016, 13, 1933-7213, 403, 10.1007/s13311-016-0429-3
    3. Vahid Hajali, Monica L. Andersen, Sajad Sahab Negah, Vahid Sheibani, Sex differences in sleep and sleep loss-induced cognitive deficits: The influence of gonadal hormones, 2019, 108, 0018506X, 50, 10.1016/j.yhbeh.2018.12.013
    4. Mariam M. Morchiladze, Tamila K. Silagadze, Zurab K. Silagadze, Visceral theory of sleep and origins of mental disorders, 2018, 120, 03069877, 22, 10.1016/j.mehy.2018.07.023
    5. Kristina Arion, Natasha L. Orr, Heather Noga, Catherine Allaire, Christina Williams, Mohamed A. Bedaiwy, Paul J. Yong, A Quantitative Analysis of Sleep Quality in Women with Endometriosis, 2020, 29, 1540-9996, 1209, 10.1089/jwh.2019.8008
    6. Serge Ade, Thierry Adoukonou, Maurice A. Badjagou, Prudence A. Wachinou, Adebayo C. Alassani, Gildas Agodokpessi, Anthony D. Harries, Sleep-related disorders and sleep quality among adults living in Parakou, a sub-Saharan African city, 2021, 1520-9512, 10.1007/s11325-021-02306-2
    7. Tyler B. Mason, Allison Engwall, Michael P. Mead, Leah A. Irish, Sleep and eating disorders among adults enrolled in a commercial weight loss program: associations with self-report and objective sleep measures, 2019, 24, 1590-1262, 307, 10.1007/s40519-019-00664-1
    8. Naima Laharnar, Stefanie Uibel, Corin Hild, Martin Glos, Thomas Penzel, Ingo Fietze, James Andrew Rowley, Overnight polysomnography and the recording of sleep and sleep-related respiration in orchestra musicians – possible protective effects of wind instruments on respiration, 2020, 15, 1932-6203, e0231549, 10.1371/journal.pone.0231549
    9. Konstantina Skolariki, Julie Seibt, 2021, Chapter 56-1, 978-3-319-75479-6, 1, 10.1007/978-3-319-75479-6_56-1
    10. Emerson M. Wickwire, David M. Schnyer, Anne Germain, Michael T. Smith, Scott G. Williams, Christopher J. Lettieri, Ashlee B. McKeon, Steven M. Scharf, Ryan Stocker, Jennifer Albrecht, Neeraj Badjatia, Amy J. Markowitz, Geoffrey T. Manley, Sleep, Sleep Disorders, and Circadian Health following Mild Traumatic Brain Injury in Adults: Review and Research Agenda, 2018, 35, 0897-7151, 2615, 10.1089/neu.2017.5243
    11. Royette Tavernier, Grant C. Hill, Tamare V. Adrien, Be well, sleep well: An examination of directionality between basic psychological needs and subjective sleep among emerging adults at university, 2019, 5, 23527218, 288, 10.1016/j.sleh.2019.02.007
    12. Yujie Sun, Na Zhang, Yuxia Qu, Yijia Cao, Jianhao Li, Yuwei Yang, Tiange Yang, Yikun Sun, Shuangxia decoction alleviates p-chlorophenylalanine induced insomnia through the modification of serotonergic and immune system, 2020, 35, 0885-7490, 315, 10.1007/s11011-019-00499-w
    13. Afifah Nawi, Koh Liew Eu, Asma Nadia Ahmad Faris, Wan Amir Nizam Wan Ahmad, Liza Noordin, Lipid peroxidation in the descending thoracic aorta of rats deprived of REM sleep using the inverted flowerpot technique, 2020, 105, 0958-0670, 1223, 10.1113/EP088667
    14. Olivia J Veatch, Brendan T Keenan, Philip R Gehrman, Beth A Malow, Allan I Pack, Pleiotropic genetic effects influencing sleep and neurological disorders, 2017, 16, 14744422, 158, 10.1016/S1474-4422(16)30339-8
    15. Carolyn E. Jones, Miranda M. Lim, Phasic Sleep Events Shape Cognitive Function after Traumatic Brain Injury: Implications for the Study of Sleep in Neurodevelopmental Disorders, 2016, 3, 2373-7972, 232, 10.3934/Neuroscience.2016.2.232
    16. Kyra Hamilton, Hei Tung Heather Ng, Chun-Qing Zhang, Daniel J. Phipps, Ru Zhang, Social Psychological Predictors of Sleep Hygiene Behaviors in Australian and Hong Kong University Students, 2020, 1070-5503, 10.1007/s12529-020-09859-8
    17. Olga Berdina, Irina Madaeva, Svetlana Bolshakova, Vladimir Polyakov, Olga Bugun, Liubov Rychkova, Alteration of sleep homeostasis and cognitive impairment in apneic obese adolescents, 2021, 1446-9235, 10.1007/s41105-021-00317-w
    18. Lily Yuen Wah Ho, Claudia Kam Yuk Lai, Shamay Sheung Mei Ng, Contribution of sleep quality to fatigue following a stroke: a cross-sectional study, 2021, 21, 1471-2377, 10.1186/s12883-021-02174-z
    19. Charu T. Tuladhar, Sophie Schwartz, Ashley M. St. John, Jerrold S. Meyer, Amanda R. Tarullo, Infant diurnal cortisol predicts sleep, 2021, 0962-1105, 10.1111/jsr.13357
    20. Sara Forslind, Harry J. Blokhuis, Anja B. Riber, Disturbance of resting behaviour of broilers under different environmental conditions, 2021, 242, 01681591, 105425, 10.1016/j.applanim.2021.105425
    21. Omokehinde Olubunmi Fakorede, Peter Olutunde Onifade, Oluyinka Emmanuel Majekodunmi, Adegboyega Ogunwale, Adefunke Omosefe DadeMatthews, Insomnia and quality of life of prisoners at a Nigerian maximum-security prison, 2021, 17, 1744-9200, 574, 10.1108/IJPH-12-2020-0097
    22. Asif iqbal Bhat, Dr Manmeet Singh, Dr Raghav Arora, Dr sunny Babber, Dr Rohit Jasrotia, Dr Parul Gupta, Prevalance of Insomnia Among Elderly Patients Attending Teritiary Care Hospital , 2022, 25, 2319-2062, 35, 10.55229/ijbs2022.0307
    23. Meta Lavrič, Vita Štukovnik, The impact of shift work on sleep and mental health – a review of the literature and a look ahead, 2021, 23505141, 153, 10.20419/2021.30.541
    24. David M. Powell, Eli Baskir, Behavior and Habitat Use Remain Diverse and Variable in Modern Zoological Exhibits over the Long-Term: Case Studies in 5 Species of Ursidae, 2021, 2, 2673-5636, 677, 10.3390/jzbg2040049
    25. Erin A. Almklov, Guadalupe L. Rivera, Henry Orff, 2022, 9780128228722, 265, 10.1016/B978-0-12-822872-2.00009-8
    26. Javier Gomez-Pilar, Gonzalo C. Gutiérrez-Tobal, Roberto Hornero, 2022, Chapter 2, 978-3-031-06412-8, 17, 10.1007/978-3-031-06413-5_2
    27. Emmanuel Frimpong, Melodee Mograss, Tehila Zvionow, Thien Thanh Dang-Vu, The effects of evening high-intensity exercise on sleep in healthy adults: A systematic review and meta-analysis, 2021, 60, 10870792, 101535, 10.1016/j.smrv.2021.101535
    28. Tian Yue, Xuefang Liu, Qi Gao, Yan Wang, Different Intensities of Evening Exercise on Sleep in Healthy Adults: A Systematic Review and Network Meta-Analysis, 2022, Volume 14, 1179-1608, 2157, 10.2147/NSS.S388863
    29. Regina Müller, Eva Kuhn, Robert Ranisch, Jonathan Hunger, Nadia Primc, Ethics of sleep tracking: techno-ethical particularities of consumer-led sleep-tracking with a focus on medicalization, vulnerability, and relationality, 2023, 25, 1388-1957, 10.1007/s10676-023-09677-y
    30. Byung Wook Song, Hye-Jin Jeong, Bo Young Kim, Yong Won Cho, Chang-Nam Son, Sung-Soo Kim, Sang-Hyon Kim, Bath Ankylosing Spondylitis Disease Activity Index is Associated With the Quality of Sleep in Ankylosing Spondylitis Patients, 2021, 28, 2093-940X, 143, 10.4078/jrd.2021.28.3.143
    31. Kinga Grochowalska, Marcin Ziętkiewicz, Ewa Więsik-Szewczyk, Aleksandra Matyja-Bednarczyk, Katarzyna Napiórkowska-Baran, Katarzyna Nowicka-Sauer, Adam Hajduk, Dariusz Sołdacki, Zbigniew Zdrojewski, Subjective sleep quality and fatigue assessment in Polish adult patients with primary immunodeficiencies: A pilot study, 2023, 13, 1664-3224, 10.3389/fimmu.2022.1028890
    32. Jennifer Sandermann, Stephanie Kurzenhäuser-Carstens, 2022, Chapter 6, 978-3-658-35830-3, 143, 10.1007/978-3-658-35831-0_6
    33. Tracy D. Hecht, Heather Cluley, Alexandru M. Lefter, Onn Anong Ngamwattana, A dynamic framework of boundary permeability: daily events and within-individual fluctuations in daily work and nonwork boundary permeation, 2022, 1359-432X, 1, 10.1080/1359432X.2022.2081075
    34. Sara Forslind, Carlos E. Hernandez, Anja B. Riber, Helena Wall, Harry J. Blokhuis, Resting behavior of broilers reared with or without artificial brooders, 2022, 9, 2297-1769, 10.3389/fvets.2022.908196
    35. Amanda Bulman, Nathan M. D’Cunha, Wolfgang Marx, Andrew J. McKune, Rati Jani, Nenad Naumovski, Nutraceuticals as Potential Targets for the Development of a Functional Beverage for Improving Sleep Quality, 2021, 7, 2306-5710, 33, 10.3390/beverages7020033
    36. Samantha S. Y. Lee, Vinay K. Nilagiri, David A. Mackey, Sleep and eye disease: A review, 2022, 50, 1442-6404, 334, 10.1111/ceo.14071
    37. Morenikeji Adeyoyin Komolafe, Oluwatosin Eunice Olorunmoteni, Kikelomo Adebanke Kolawole, Olufemi K. Ogundipe, Michael Bimbola Fawale, Akintunde Adeolu Adebowale, Ahmed Omokayode Idowu, Ahmad Abefe Sanusi, Josephine Eniola A. Eziyi, Kolawole Samuel Mosaku, 2023, 9789815049367, 229, 10.2174/9789815049367123010021
    38. A. Mondino, C. Ludwig, C. Menchaca, K. Russell, K. E. Simon, E. Griffith, A. Kis, B. D. X. Lascelles, M. E. Gruen, N. J. Olby, Development and validation of a sleep questionnaire, SNoRE 3.0, to evaluate sleep in companion dogs, 2023, 13, 2045-2322, 10.1038/s41598-023-40048-1
    39. Elisa Perger, Laura Calvillo, Riccardo Cremascoli, 2024, Chapter 16, 978-3-031-18544-1, 191, 10.1007/978-3-031-18545-8_16
    40. Zhong‐Wen Jiang, Liang Ma, Shi‐ang Tao, Cheng Wenda, Chuyu Cheng, Dan‐yang Wu, Wei‐Guo Du, Analysis of resting status reveals distinct elevational variation in metabolisms of lizards, 2024, 105, 0012-9658, 10.1002/ecy.4414
    41. Sergio Garbarino, Nicola Luigi Bragazzi, Revolutionizing Sleep Health: The Emergence and Impact of Personalized Sleep Medicine, 2024, 14, 2075-4426, 598, 10.3390/jpm14060598
    42. Konstantina Skolariki, Julie Seibt, 2023, Chapter 56, 978-3-319-75921-0, 427, 10.1007/978-3-319-75922-7_56
    43. Bian Ma, Lijuan Duan, Zhaoyang Huang, Yuanhua Qiao, Bei Gong, A federated semi-supervised automatic sleep staging method based on relationship knowledge sharing, 2024, 237, 09574174, 121427, 10.1016/j.eswa.2023.121427
    44. Madeleine J. Smith, Michael Pellegrini, Brendan Major, Marnie Graco, Stephanie Porter, Sharon Kramer, Katherine Sewell, Sabrina Salberg, Zhibin Chen, Richelle Mychasiuk, Natasha A. Lannin, Improving physical movement during stroke rehabilitation: investigating associations between sleep measured by wearable actigraphy technology, fatigue, and key biomarkers, 2024, 21, 1743-0003, 10.1186/s12984-024-01380-3
    45. Endre Putyora, Sarah Brocklehurst, Victoria Sandilands, The Effects of Commercially-Relevant Disturbances on Sleep Behaviour in Laying Hens, 2023, 13, 2076-2615, 3105, 10.3390/ani13193105
    46. Patricia Frytz, Dominik P. J. Heib, Kerstin Hoedlmoser, Soccer, Sleep, Repeat: Effects of Training Characteristics on Sleep Quantity and Sleep Architecture, 2023, 13, 2075-1729, 1679, 10.3390/life13081679
    47. Vanessa Giffoni M. N. P. Peixoto, Lucas Alves Facci, Thiago C. S. Barbalho, Raíssa Nascimento Souza, Alice Mendes Duarte, Katie Moraes Almondes, The context of COVID-19 affected the long-term sleep quality of older adults more than SARS-CoV-2 infection, 2024, 15, 1664-0640, 10.3389/fpsyt.2024.1305945
    48. Barbara Grabowska, Joanna Żminda, Luba Ślósarz, Physical activity and sleep quality of young people – formation of habits in the family of origin, 2023, 30, 2082-9019, 223, 10.61905/wwr/176125
    49. Andrew Wooyoung Kim, Psychosocial stress, adult suicidal ideation, and the mediating effect of poor sleep quality during the COVID‐19 pandemic in South Africa, 2024, 36, 1042-0533, 10.1002/ajhb.24038
    50. Halil Dogrul, Yusuf Cetin Doganer, Umit Aydogan, Rumeysa Nur Bayrak, Association Between Sleep Hygiene Parameters and Sleep Habits in 5- to 10-Year-Old School-Age Children, 2024, 0009-9228, 10.1177/00099228241283276
    51. Amanda Azevedo de Carvalho, Dante Ogassavara, Thais da Silva-Ferreira, Patricia Costa Lima Tierno, Jeniffer Ferreira-Costa, José Maria Montiel, MANUTENÇÃO DA HOMEOSTASE E DO FUNCIONAMENTO BIOLÓGICO: Implicações do dormir e do sono no cotidiano das pessoas, 2024, 10, 2446922X, 485, 10.22289/2446-922X.V10N1A30
    52. Hannah Scott, Bastien Lechat, Alisha Guyett, Amy C. Reynolds, Nicole Lovato, Ganesh Naik, Sarah Appleton, Robert Adams, Pierre Escourrou, Peter Catcheside, Danny J. Eckert, Sleep Irregularity Is Associated With Hypertension: Findings From Over 2 Million Nights With a Large Global Population Sample, 2023, 80, 0194-911X, 1117, 10.1161/HYPERTENSIONAHA.122.20513
    53. Vinícius Rosa Cota, Simone Del Corso, Gianluca Federici, Gabriele Arnulfo, Michela Chiappalone, Efficient Sleep–Wake Cycle Staging via Phase–Amplitude Coupling Pattern Classification, 2024, 14, 2076-3417, 5816, 10.3390/app14135816
    54. Isabelle Ross, Leigh Signal, Natasha Tassell-Matamua, Robert Meadows, Rosemary Gibson, Sleep as a social and cultural practice in Aotearoa: a scoping review, 2024, 1177-083X, 1, 10.1080/1177083X.2024.2403654
    55. Nils Runge, Melanie Suckow, Die Rolle von Schlafproblemen in der muskuloskelettalen Physiotherapie – Teil 1, 2023, 27, 2701-6986, 100, 10.1055/a-2013-4589
    56. Marjolein Meijdam, Marcel Eens, Wendt Müller, Artificial light at night impairs inhibitory control in a wild songbird, 2023, 885, 00489697, 163765, 10.1016/j.scitotenv.2023.163765
    57. Graham Joseph Adams, Philip A. O'Brien, The unified theory of sleep: Eukaryotes endosymbiotic relationship with mitochondria and REM the push-back response for awakening, 2023, 15, 24519944, 100100, 10.1016/j.nbscr.2023.100100
    58. Mary Breheny, Isabelle Ross, Clare Ladyman, Leigh Signal, Kevin Dew, Rosemary Gibson, Stephen Katz, “It’s Just [Complicated] Sleep”: Discourses of Sleep and Aging in the Media, 2023, 63, 0016-9013, 1591, 10.1093/geront/gnad058
    59. Kristine M. Gandia, Elizabeth S. Herrelko, Sharon E. Kessler, Hannah M. Buchanan-Smith, Understanding Circadian and Circannual Behavioral Cycles of Captive Giant Pandas (Ailuropoda melanoleuca) Can Help to Promote Good Welfare, 2023, 13, 2076-2615, 2401, 10.3390/ani13152401
    60. Rita Ferreira, João Brás, Joana Fialho, Cristina Peixoto, Sleep Quality among Medical Students of a Portuguese University, 2024, 10, 2184-5417, 5, 10.51338/rppsm.496
    61. Petra Höferová, Eva Aigelová, Reasons and impacts of hypnotic overuse in seniors, 2024, 25, 12130508, E1, 10.36290/psy.2024.026
    62. Bassem Khalfi, Kobe Buyse, Imad Khan, Camila Lopes Carvalho, Patricia Soster, Gunther Antonissen, Frank André Maurice Tuyttens, Cooled Multifunctional Platforms to Alleviate Heat Stress in Broiler Chickens: Effects on Performance, Carcass and Meat Quality Traits, 2024, 14, 2076-2615, 3448, 10.3390/ani14233448
    63. F. Gallo, A. Myachykov, J. Abutalebi, V. DeLuca, J. Ellis, J. Rothman, L.R. Wheeldon, Bilingualism, sleep, and cognition: An integrative view and open research questions, 2025, 260, 0093934X, 105507, 10.1016/j.bandl.2024.105507
    64. Lieke L. ten Brummelhuis, Charles Calderwood, Christopher C. Rosen, Allison S. Gabriel, Peaking Today, Taking It Easy Tomorrow: Daily Performance Dynamics of Working Long Hours, 2024, 0894-3796, 10.1002/job.2847
    65. Christos Sikaras, Argyro Pachi, Sofia Alikanioti, Ioannis Ilias, Eleni Paraskevi Sideri, Athanasios Tselebis, Aspasia Panagiotou, Occupational Burnout and Insomnia in Relation to Psychological Resilience Among Greek Nurses in the Post-Pandemic Era, 2025, 15, 2076-328X, 126, 10.3390/bs15020126
    66. Carolin A.B. Adler, 2025, 9780128096338, 10.1016/B978-0-443-29068-8.00081-7
    67. Hannah Scott, Michael Perlis, The Sleep Opportunity, Need and Ability (SONA) Theory, 2025, 0962-1105, 10.1111/jsr.70030
    68. Dawit Afewerki Meles, Ashenafi Damte Ayele, Hagos Tsegaberhan, Tilahun Belete Mossie, Quality of sleep and associated factors among medical interns in public universities in North Ethiopia, 2025, 16, 1664-0640, 10.3389/fpsyt.2025.1448028
    69. Daniel H. Cooper, Isaac Almendros, Tetyana Kendzerska, Sleep, Circadian Rhythms, and Lung Cancer, 2025, 1069-3424, 10.1055/a-2531-1059
    70. Joana Belo, Miguel Meira Cruz, Carla Viegas, Joana Lage, Susana Marta Almeida, Sandra Cabo Verde, Célia Alves, Nuno Canha, Sleep and indoor air quality: an exploratory polysomnographic evaluation of potential associations, 2025, 13091042, 102557, 10.1016/j.apr.2025.102557
  • Reader Comments
  • © 2020 the Author(s), licensee AIMS Press. This is an open access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0)
通讯作者: 陈斌, bchen63@163.com
  • 1. 

    沈阳化工大学材料科学与工程学院 沈阳 110142

  1. 本站搜索
  2. 百度学术搜索
  3. 万方数据库搜索
  4. CNKI搜索

Metrics

Article views(6167) PDF downloads(268) Cited by(2)

Figures and Tables

Figures(4)  /  Tables(3)

/

DownLoad:  Full-Size Img  PowerPoint
Return
Return

Catalog