Estimating the effect of a predictor measured by two informants on a continuous outcome: a comparison of methods.
ABSTRACT: Investigators sometimes use information obtained from multiple informants about a given variable. We focus on estimating the effect of a predictor on a continuous outcome, when that (true) predictor cannot be observed directly but is measured by 2 informants. We describe various approaches to using information from 2 informants to estimate a regression or correlation coefficient for the effect of the (true) predictor on the outcome. These approaches include methods we refer to as single informant, simple average, optimal weighted average, principal components analysis, and classical measurement error. Each of these 5 methods effectively uses a weighted average of the informants' reports as a proxy for the true predictor in calculating the correlation or regression coefficient. We compare the performance of these methods in simulation experiments that assume a rounded congeneric measurement model for the relationship between the informants' reports and a true predictor that is a mixture of zeros and positively distributed continuous values. We also compare the methods' performance in a real data example-the relationship between vigorous physical activity (the predictor) and body mass index (the continuous outcome). The results of the simulations and the example suggest that the simple average is a reasonable choice when there are only 2 informants.
Project description:Methods for multiple informants help to estimate the marginal effect of each multiple source predictor and formally compare the strength of their association with an outcome. We extend multiple informant methods to the case of hierarchical data structures to account for within cluster correlation. We apply the proposed method to examine the relationship between features of the food environment near schools and children's body mass index z-scores (BMIz). Specifically, we compare the associations between two different features of the food environment (fast food restaurants and convenience stores) with BMIz and investigate how the association between the number of fast food restaurants or convenience stores and child's BMIz varies across distance from a school. The newly developed methodology enhances the types of research questions that can be asked by investigators studying effects of environment on childhood obesity and can be applied to other fields.
Project description:This study examined how children's intuitions and informants' expertise influence children's trust in informants' claims. Three- to 8-year-olds (N = 192) watched videos in which experts (animal/biology experts or artifact/physics experts) made either intuitively plausible or counterintuitive claims about obscure animals or artifacts. Claims fell either within or beyond experts' domains of expertise. Children of all ages were more trusting of claims made by informants with relevant, as opposed to irrelevant, expertise. Children also showed greater acceptance of intuitive rather than counterintuitive claims, a differentiation that increased with age as they developed firmer intuitions about what can ordinarily happen. In summary, children's trust in testimony depends on whether informants have the relevant expertise as well as on children's own developing intuitions.
Project description:This paper presents the results of a survey among key informants that was conducted between June and September 2011 in Breast Cancer Centers that were accredited according to the criteria of the German Cancer Society (DKG). The survey intended to assess the degree to which the breast cancer center concept was accepted among the key informants as well as to gain an overview over structures and processes in the centers. The Questionnaire for Breast Cancer Centres Key Informants 2011 (FRIZ 2011) was used with two reminders having been sent out. Questionnaires were sent back from 149 of the 243 initially contacted hospitals (response rate: 61.3?%). The vast majority of respondents indicated to be part of the Breast Cancer Center management. 110 of the 149 hospitals did also participate in the patient survey conducted in 2010. Among the key informants surveyed, the concept is highly accepted with regard to improvements in patient care. Overall, the concept is regarded as "good" or "very good" by almost all respondents. Both contact to resident doctors and the hospitals' reputations improved since the implementation of the concept. Quality and patient safety were more often on the agenda than financial performance in the quality circles with the main co-operation partners of the Breast Cancer Centers.
Project description:The idea of what is natural has particular relevance in the thoroughbred racing and breeding discourse. It guides breeding regulations; influences how the thoroughbreds' behaviour is perceived and has implications for husbandry, handling, training and racing practices. This study investigates how key industry and animal advocacy informants based in the US, Australia and the UK conceptualise naturalness within the context of common racing practices that potentially impact the horses' welfare. The informants were interviewed using semi-structured interviewing and photo-elicitation. Four common images of thoroughbreds on race day were presented to elicit the informants' responses. Differences emerged between how the two groups tended to describe the images and the role naturalness played in their conceptualisations. The findings were analysed using an updated version of the Layers of Engagement with Animal Protection developed by Bergmann to situate the informants' conceptualisations of naturalness within the wider thoroughbred protection discourse. In conclusion, the industry informants tended to defend the status quo of common racing practices. They tended to naturalise and normalise these practices and downplay their welfare impact. This poses risks for thoroughbred welfare, which are amplified by misrepresentations of what is natural. With the public's understanding of welfare and racing practices growing, racing's legitimacy may be further questioned. Opportunities to leverage the potential of the notion of naturalness for thoroughbred protection are discussed.
Project description:Imputation strategies are widely used in settings that involve inference with incomplete data. However, implementation of a particular approach always rests on assumptions, and subtle distinctions between methods can have an impact on subsequent analyses. In this research article, we are concerned with regression models in which the true underlying relationship includes interaction terms. We focus in particular on a linear model with one fully observed continuous predictor, a second partially observed continuous predictor, and their interaction. We derive the conditional distribution of the missing covariate and interaction term given the observed covariate and the outcome variable, and examine the performance of a multiple imputation procedure based on this distribution. We also investigate several alternative procedures that can be implemented by adapting multivariate normal multiple imputation software in ways that might be expected to perform well despite incompatibilities between model assumptions and true underlying relationships among the variables. The methods are compared in terms of bias, coverage, and CI width. As expected, the procedure based on the correct conditional distribution performs well across all scenarios. Just as importantly for general practitioners, several of the approaches based on multivariate normality perform comparably with the correct conditional distribution in a number of circumstances, although interestingly, procedures that seek to preserve the multiplicative relationship between the interaction term and the main-effects are found to be substantially less reliable. For illustration, the various procedures are applied to an analysis of post-traumatic stress disorder symptoms in a study of childhood trauma.
Project description:This paper employs an un-weighted and weighted exergy network to study the properties of ferrous metal ores in countries worldwide and their evolution from 2002 to 2012. We find that there are few countries controlling most of the ferrous metal ore exports in terms of exergy and that the entire exergy flow network is becoming more heterogeneous though the addition of new nodes. The increasing of the average clustering coefficient indicates that the formation of an international exergy flow system and regional integration is improving. When we contrast the average out strength of exergy and the average out strength of currency, we find both similarities and differences. Prices are affected largely by human factors; thus, the growth rate of the average out strength of currency has fluctuated acutely in the eleven years from 2002 to 2012. Exergy is defined as the maximum work that can be extracted from a system and can reflect the true cost in the world, and this parameter fluctuates much less. Performing an analysis based on the two aspects of exergy and currency, we find that the network is becoming uneven.
Project description:Background:Hyperglycaemia is associated with a worse outcome in acute ischaemic stroke patients; yet the pathophysiological mechanisms of hyperglycaemia-induced damage are poorly understood. We hypothesised that hyperglycaemia at the time of stroke onset exacerbates ischaemic brain damage by increasing the severity of the blood flow deficit. Methods:Adult, male Wistar rats were randomly assigned to receive vehicle or glucose solutions prior to permanent middle cerebral artery occlusion. Cerebral blood flow was assessed semi-quantitatively either 1?h after middle cerebral artery occlusion using 99mTc-D, L-hexamethylpropyleneamine oxime (99mTc-HMPAO) autoradiography or, in a separate study, using quantitative pseudo-continuous arterial spin labelling for 4?h after middle cerebral artery occlusion. Diffusion weighted imaging was performed alongside pseudo-continuous arterial spin labelling and acute lesion volumes calculated from apparent diffusion coefficient maps. Infarct volume was measured at 24?h using rapid acquisition with refocused echoes T2-weighted magnetic resonance imaging. Results:Glucose administration had no effect on the severity of ischaemia when assessed by either 99mTc-HMPAO autoradiography or pseudo-continuous arterial spin labelling perfusion imaging. In comparison to the vehicle group, apparent diffusion coefficient-derived lesion volume 2-4?h post-middle cerebral artery occlusion and infarct volume 24?h post-middle cerebral artery occlusion were significantly greater in the glucose group. Conclusions:Hyperglycaemia increased acute lesion and infarct volumes but there was no evidence that the acute blood flow deficit was exacerbated. The data reinforce the conclusion that the detrimental effects of hyperglycaemia are rapid, and that treatment of post-stroke hyperglycaemia in the acute period is essential but the mechanisms of hyperglycaemia-induced harm remain unclear.
Project description:BACKGROUND: In medical practice many, essentially continuous, clinical parameters tend to be categorised by physicians for ease of decision-making. Indeed, categorisation is a common practice both in medical research and in the development of clinical prediction rules, particularly where the ensuing models are to be applied in daily clinical practice to support clinicians in the decision-making process. Since the number of categories into which a continuous predictor must be categorised depends partly on the relationship between the predictor and the outcome, the need for more than two categories must be borne in mind. METHODS: We propose a categorisation methodology for clinical-prediction models, using Generalised Additive Models (GAMs) with P-spline smoothers to determine the relationship between the continuous predictor and the outcome. The proposed method consists of creating at least one average-risk category along with high- and low-risk categories based on the GAM smooth function. We applied this methodology to a prospective cohort of patients with exacerbated chronic obstructive pulmonary disease. The predictors selected were respiratory rate and partial pressure of carbon dioxide in the blood (PCO2), and the response variable was poor evolution. An additive logistic regression model was used to show the relationship between the covariates and the dichotomous response variable. The proposed categorisation was compared to the continuous predictor as the best option, using the AIC and AUC evaluation parameters. The sample was divided into a derivation (60%) and validation (40%) samples. The first was used to obtain the cut points while the second was used to validate the proposed methodology. RESULTS: The three-category proposal for the respiratory rate was ??20;(20,24];>?24, for which the following values were obtained: AIC=314.5 and AUC=0.638. The respective values for the continuous predictor were AIC=317.1 and AUC=0.634, with no statistically significant differences being found between the two AUCs (p?=0.079). The four-category proposal for PCO2 was ??43;(43,52];(52,65];>?65, for which the following values were obtained: AIC=258.1 and AUC=0.81. No statistically significant differences were found between the AUC of the four-category option and that of the continuous predictor, which yielded an AIC of 250.3 and an AUC of 0.825 (p?=0.115). CONCLUSIONS: Our proposed method provides clinicians with the number and location of cut points for categorising variables, and performs as successfully as the original continuous predictor when it comes to developing clinical prediction rules.
Project description:OBJECTIVES:Pretreatment identification of radio-insensitive head and neck squamous cell carcinomas (HNSCC) would affect treatment modality selection. The apparent diffusion coefficient (ADC) of a tumor could be a predictor of local recurrence. However, little is known about its prognostic value next to known factors such as clinical T-stage. The aim of the present study is to determine the added value of pretreatment ADC to clinical T-stage as a prognostic factor for local recurrence. METHODS:This retrospective cohort study included 217 patients with HNSCC treated with (chemo)radiotherapy between April 2009 and December 2015. All patients underwent diffusion-weighted MRI prior to treatment. Median ADC values of all tumors were obtained using a semi-automatic delineation method. Univariate models containing ADC and T-stage were compared with a multivariable model containing both variables. RESULTS:Fifty-eight patients experienced a local recurrence within 3 years. On average, the ADC value in the group of patients with a recurrence was 1.01 versus 1.00 (10-3 mm2/s) in the group without a recurrence. Univariate analysis showed no significant association between tumor ADC and local recurrence within 3 years after (chemo)radiotherapy (p?=?0.09). Cox regression showed that clinical T-stage was an independent predictor of local recurrence and adding ADC to the model did not increase its performance. CONCLUSION:Pretreatment ADC has no added value as a prognostic factor for local recurrence to clinical T-stage. KEY POINTS:• Pretreatment identification of head and neck squamous cell carcinoma patients who do not benefit from (chemo)radiotherapy could improve personalized cancer care. • The apparent diffusion coefficient (ADC) obtained from diffusion-weighted MRI has been reported to be a prognostic factor for local recurrence. • In this study, ADC has no added value as a prognostic factor compared with clinical T-stage.
Project description:Background:The Penn Parkinson's Daily Activities Questionnaire-15 (PDAQ-15) assesses cognition-related instrumental activities of daily living (IADL) in Parkinson's disease (PD). Objectives:To assess the degree and predictors of disagreement between patients (PT) and knowledgeable informants (KI) on the PDAQ-15. Methods:We recruited 254 PT and KI pairs (PT-KI), determined predictors of agreement, and compared scores to a performance-based functional measure (Direct Assessment of Functional Status [DAFS]; N = 61). Results:PT and KI total score (intraclass correlation = 0.57) and individual item (Cohen's kappa = 0.46-0.62) agreement were moderate. Patient depression, global cognition, and caregiver burden (all P < 0.05), predicted PT-KI discrepancy. PT-KI discrepancy was highest in patients with a dementia diagnosis, followed by mild cognitive impairment and then normal cognition (all P < 0.01), with PT rating themselves relatively more functionally intact as cognition worsened. DAFS performance was more highly correlated with KI (r = 0.82; P < 0.001) than PT (r = 0.62; P < 0.001) PDAQ-15 score. Conclusions:Our results support using KI as proxies when assessing cognitive IADLs in PD PTs, particularly in cases of more advanced cognitive decline.