离散事件模拟模型在流行病学筛查成本效果分析中的应用

离散事件模拟(DES)模型基于个体水平的数据, 通过模拟发生在不同时间的各离散事件的情况来反映疾病发生发展过程, 不仅可以处理个体特征对疾病进展的影响, 由于其不设定固定的周期, 对事件发生时间的估计相比状态转换模型更精确, 特别适用于考虑时间维度的复杂疾病筛查及干预策略的长期效果评价。本文介绍了DES模型的基本原理、构建步骤、分析方法及相关注意事项。结合在英国≥65岁的女性人群中开展的一项腹主动脉瘤筛查成本效果分析的研究实例, 从模型构建、模型分析和结果解读等方面, 详细讨论了DES模型在复杂疾病筛查成本效果分析中实际应用的要点。DES模型通过建立离散事件发生时间的分布函数, 实现对事件发生时间的精确估计, 越来越多地应用于需考虑时间维度的复杂疾病筛查策略评价。在建模过程中应注意清晰展现模型结构和模拟过程, 并遵循相应的报告规范开展成本效果分析, 以保证研究的透明度和可重复性。     

Insurees’ preferences in hospital choice—A population-based study

In Germany, the patient himself makes the choice for or against a health service provider. Hospital comparison websites offer him possibilities to inform himself before choosing. However, it remains unclear, how health care consumers use those websites, and there is little information about how preferences in hospital choice differ interpersonally.We conducted a Discrete-Choice-Experiment (DCE) on hospital choice with 1500 randomly selected participants (age 40–70) in three different German cities selecting four attributes for hospital vignettes. The analysis of the study draws on multilevel mixed effects logit regression analyses with the dependent variables: “chance to select a hospital” and “choice confidence”. Subsequently, we performed a Latent-Class-Analysis to uncover consumer segments with distinct preferences.590 of the questionnaires were evaluable. All four attributes of the hospital vignettes have a significant impact on hospital choice. The attribute “complication rate” exerts the highest impact on consumers’ decisions and reported choice confidence. Latent-Class-Analysis results in one dominant consumer segment that considered the complication rate the most important decision criterion.Using DCE, we were able to show that the complication rate is an important trusted criterion in hospital choice to a large group of consumers. Our study supports current governmental efforts in Germany to concentrate the provision of specialized health care services. We suggest further national and cross-national research on the topic.     

The effects of convenience and quality on the demand for vaccination: Results from a discrete choice experiment

BackgroundVaccination is an effective way to prevent infectious diseases. Most studies analysed people's vaccine decisions, but few studies have analysed the effects of convenience such as immunisation schedule and distance and the quality of vaccination service on vaccination uptake.ObjectiveThe aim of this paper was to investigate adults’ preferences for convenience and quality of vaccination service, calculate the private economic benefit from convenience (vaccination schedule and distance) and quality, and predict the uptake rate for different vaccine scenarios.MethodsIn our study, we interviewed 266 adults in 2 counties of Shandong province in China. The discrete choice experiment (DCE) was employed to analyse the preference for hepatitis B virus (HBV) vaccination, and a mixed logit model was used to estimate respondent preferences for vaccination attributes included in the DCE.ResultsThe protection rate against hepatitis B (HB), duration of protection, risk of side effects, vaccination cost, schedule, and vaccination sites were proved to influence adults’ preferences for HBV vaccination. The estimated willingness to pay (WTP) for 1 dose schedule instead of 3 doses and for a third-level vaccination site instead of a first-level site was almost equal (19 RMB). However, if the protection duration of the vaccination programme changed from 5 years to 20 years, the adults were willing to pay 35.05 RMB, and WTP for a 99% protection rate instead of a 79% rate was 67.71 RMB. The predicted uptake rate is almost 43% for the base case of HBV vaccination.ConclusionsAdults made trade-offs between vaccination schedules, vaccination sites, and other characteristics of HBV vaccine. The impact of attributes of the vaccine itself, especially protection rate against HB, duration of protection, and risk of side-effects, is more dramatic than convenience and quality of vaccination service.     

The influence of gene expression profiling on decisional conflict in decision making for early-stage breast cancer chemotherapy

BackgroundWomen with early-stage breast cancer, of whom only 15% will experience a recurrence, are often conflicted or uncertain about taking chemotherapy. Gene expression profiling (GEP) of tumours informs risk prediction, potentially affecting treatment decisions. We examined whether receiving a GEP test score reduces decisional conflict in chemotherapy treatment decision making.MethodsA general population sample of 200 women completed the decisional conflict scale (DCS) at baseline (no GEP test score scenario) and after (scenario with GEP test score added) completing a discrete choice experiment survey for early-stage breast cancer chemotherapy. We scaled the 16-item DCS total scores and subscores from 0 to 100 and calculated means, standard deviations and change in scores, with significance (p < 0.05) based on matched pairs t-tests.ResultsWe identified five respondent subgroups based on preferred treatment option; almost 40% did not change their chemotherapy decision after receiving GEP testing information. Total score and all subscores (uncertainty, informed, values clarity, support, and effective decision) decreased significantly in the respondent subgroup who were unsure about taking chemotherapy initially but changed to no chemotherapy (n =33). In the subgroup of respondents (n = 25) who chose chemotherapy initially but changed to unsure, effective decision subscore increased significantly. In the overall sample, changes in total and all subscores were non-significant.ConclusionsGEP testing adds value for women initially unsure about chemotherapy treatment with a decrease in decisional conflict. However, for women who are confident about their treatment decisions, GEP testing may not add value. Decisions to request GEP testing should be personalised based on patient preferences.     

Temporally-related stimulus properties of δ9-tetrahydrocannabinol in monkeys

A two-choice operant discrimination procedure was used to train three monkeys to respond differentially between a fixed dose of ⁹-tetrahydrocannabinol ( ⁹-THC) and the drug vehicle alone. During the acquisition of the drug discrimination, both drug and vehicle alone were administered orally 2.5 h prior to the experimental session. The drug-stimulus time course was than tested by administering either the same dose of ⁹-THC or the vehicle alone from 0.5 to 16.5 h before the session. Varying administration time had no effect on the established discrimination during vehicle alone test sessions. However, orderly time-related generalization gradients were obtained around the drug administration time used during discrimination training.The authors thank Björn Jögi for his assistance in running the experiment. Research supported by NIMH Grant DA00355. Synthetic ⁹-THC obtained by approval of the FDA-NIMH Psychotomimetic Agents Advisory Committee. The animals in this study were maintained in accordance with Guide for Laboratory Animal Facilities and Care as published by the National Academy of Sciences-National Research Council.     

Preference for saccharin-sweetened alcohol relative to isocaloric sucrose

 This experiment tested the reinforcing efficacy of a saccharin-sweetened alcohol solution relative to an isocaloric sucrose drink in rats. One dipper served 10% alcohol plus 0.25% saccharin, and a second, concurrently available, dipper served 14.2% sucrose. During the course of the experiment, access to the two drinks was challenged by increasing the schedule requirement (variable-interval) that determined when a lever press would operate the dipper. There were two main findings. First, the rats continued to consume significant amounts of alcohol despite access to the isocaloric sucrose solution. Second, schedule-requirement increases that decreased sucrose-reinforced responding failed to decrease saccharin-sweetened alcohol reinforced responding. These results extend and replicate earlier findings from studies in which alcohol was mixed with sucrose, and the alcohol mixtures held a caloric advantage over the competing sucrose solutions. The experiment also included controls for differences in baseline response rates and for the influence of saccharin on preference. In the baseline response-rate control conditions, the two reinforcers were 10% sucrose and a mixture of 10% sucrose-plus-quinine. The results showed that the persistence of sweetened-alcohol reinforced responding could not be explained by differences in baseline response rates or the reinforcing properties of saccharin. Rather, the findings were consistent with the idea that the rats were defending baseline levels of alcohol-plus-saccharin consumption. Received: 15 June 1996 / Final version: 13 August 1996     

Routine Activities and Alcohol Use: Constraints on Outlet Utilization

Studies of consumers' use of alcohol beverage outlets have provided a basis for understanding drinking behaviors in different drinking environments. These studies have shown that drinking environments are related to both demographic and drinking pattern measures. Absent from these studies has been a theoretical basis on which to make predictions regarding drinking patterns and choices of drinking environments under the various social, economic, and environmental constraints typically confronting alcohol consumers. This study presents one such theoretical approach.
The approach assumes that, in the context of individual preferences for alcohol, drinking choices are constrained by consumers' economic and time-energy budgets for consumption. All other things being equal, it is suggested that greater budgets for consumption will be related to greater alcohol use, quality of beverages purchased, amenity values of purchase locations, or all three. Because on-premise drinking entails greater economic costs, greater drinking levels will be related to lower utilization of on-premise establishments.
The predictions of this approach were tested using data obtained from telephone surveys of consumers conducted in 1990 and 1991. The results showed that controlling for income, variables related to greater time-energy budgets for consumption (i.e., marital status and household composition) were related to greater consumption levels and greater utilization of on-premise establishments. Controlling for demographic measures, greater income was related to greater utilization of restaurants and increased beverage quality. Controlling for all other measures, frequencies of consumption were inversely related to consumption at on-premise establishments, reflecting the expected moderation in costs for heavier consumers on a limited alcohol budget.     

Oral self-administration of etonitazene in rhesus monkeys: Use of a fading procedure to establish etonitazene as a reinforcer

The establishment of orally delivered etonitazene (a potent opioid) as a reinforcer, was studied in eight rhesus monkeys. Initially, when given concurrent access to 2.5 μg/ml etonitazene and the water vehicle, five of the monkeys rejected the drug, whereas the other three monkeys consumed more drug solution than water. The five monkeys that rejected the drug solution underwent an acquisition phase to establish the drug as a reinforcer. A fading procedure was used to transfer control of responding from a 2% (wt/vol) ethanol solution to a 2.5 μg/ml etonitazene solution. Initially, responding was maintained by contingent deliveries of 2% ethanol. Next, across blocks of six or more sessions, increasing amounts of etonitazene were added in steps to the 2% ethanol solution. Subsequently, the 2% ethanol solution was decreased in steps to zero, leaving only the 2.5 μg/ml etonitazene present. When the fading procedure was completed, dose of etonitazene was varied by increasing the volume delivered, first under fixed ratio (FR 4) and then under an FR 8 reinforcement schedule. The same dose manipulations were made with the three monkeys who did not undergo the fading procedure because they preferred etonitazene over water when first tested. Etonitazene was established as a reinforcer for six of the eight monkeys because drug deliveries exceeded vehicle deliveries across a range of drug doses.     

Ethanol regulated preference in rats

A series of experiments evaluated the determinants of preference for mixtures of ethanol plus sucrose relative to sucrose in rats. One dipper served 10% ethanol mixed with 10% sucrose, and the second dipper served 10% sucrose. Lever presses operated each dipper according to a variable-interval 5-s schedule. In three experiments the subjects were given pre-session meals of sucrose (2.5–20 ml) or sucrose (20 ml) plus chow (5 or 10 g). Pre-session meals decreased responding maintained by sucrose but not responding maintained by ethanol mixture. In two experiments body weight was varied from 85% to 125% of the initial free-feeding values. Increases in body weight, like pre-session meals, decreased responding reinforced by sucrose, but typically did not decrease responding reinforced by ethanol mixture. Throughout most of the study, ethanol consumption remained at about 1.25 ml per half hour session (3–4 g/kg per 30 min). For example, pre-session access to ethanol mixture decreased within-session ethanol consumption, but total consumption, counting both sources, remained about 1.25 ml/session. The within-session patterns of responding also differed. Responding reinforced by ethanol mix decreased as a function of ethanol consumption, whereas responding reinforced by sucrose was relatively constant throughout the session. The simplest explanation of the results is that ethanol's pharmacological consequences regulated preference.     

Facilitation of stimulus detection performance of rats with d-amphetamine: a function of dose and level of training

 Conditions under which amphetamine may facilitate stimulus detection task choice performance in rats were investigated. Rats (n=15) were trained in a two-choice, light-detection task to three successively more stringent criterion levels of task training (minimal, intermediate, and extended) and then tested after administration of saline, 0.25, 0.50, and 0.75 mg/kg d-amphetamine (AMP). For each training level, baseline levels of choice accuracy were maintained at approximately 82% by manipulating the animals’ cue duration. No aspect of performance was enhanced by any dose of AMP after minimal criteria training, and there was a dose-dependent decrease in the number of trials completed. After the intermediate level of training, the 0.25 mg/kg dose of AMP reliably increased choice accuracy, there was no reliable change in choice reaction time, and there was a dose-dependent decrease in the number of trials completed. After the extended training, the 0.25 mg/kg dose of AMP reliably increased choice response accuracy, the 0.25 and 0.50 mg/kg doses of AMP reliably decreased choice reaction time, and there was no reliable change in the number of trials completed at any dose of AMP. These results support the contention that psychostimulants can facilitate the choice performance of rats in stimulus detection tasks if an appropriately low dose is used and the animal’s behavior is strongly controlled by the stimulus-reinforcement contingencies of the task. Received: 24 September 1997 / Final version: 9 May 1998