It has been well known that ignoring measurement error may result in substantially biased estimates in many contexts including linear and nonlinear regressions. For survival data with measurement error in covariates there has been extensive discussion in the literature with the focus being on the Cox proportional hazards models. However, the impact of measurement error on accelerated failure time (AFT) models has received little attention, though AFT models are very useful in survival data analysis. In this paper, we discuss AFT models with error-prone covariates and study the bias induced by the naive approach of ignoring measurement error in covariates. To adjust for such a bias, we describe a simulation and extrapolation method. This method is appealing because it is simple to implement and it does not require modelling the true but error-prone covariate process that is often not observable. Asymptotic normality for the resulting estimators is established. Simulation studies are carried out to evaluate the performance of the proposed method as well as the impact of ignoring measurement error in covariates. The proposed method is applied to analyse a data set arising from the Busselton Health study (Australian J. Public Health 1994; 18:129-135).
Most epidemiologic studies concerned with Major Depressive Disorder have employed cross-sectional study designs. Assessment of lifetime prevalence in such studies depends on recall of past depressive episodes. Such studies may underestimate lifetime prevalence because of incomplete recall of past episodes (recall bias). An opportunity to evaluate this issue arises with a prospective Canadian study called the National Population Health Survey (NPHS).
The NPHS is a longitudinal study that has followed a community sample representative of household residents since 1994. Follow-up interviews have been completed every two years and have incorporated the Composite International Diagnostic Interview short form for major depression. Data are currently available for seven such interview cycles spanning the time frame 1994 to 2006. In this study, cumulative prevalence was calculated by determining the proportion of respondents who had one or more major depressive episodes during this follow-up interval.
The annual prevalence of MDD ranged between 4% and 5% of the population during each assessment, consistent with existing literature. However, 19.7% of the population had at least one major depressive episode during follow-up. This included 24.2% of women and 14.2% of men. These estimates are nearly twice as high as the lifetime prevalence of major depressive episodes reported by cross-sectional studies during same time interval.
In this study, prospectively observed cumulative prevalence over a relatively brief interval of time exceeded lifetime prevalence estimates by a considerable extent. This supports the idea that lifetime prevalence estimates are vulnerable to recall bias and that existing estimates are too low for this reason.
Bias in self-reported dietary intake is important when evaluating the effect of dietary interventions, particularly for intervention foods. However, few have investigated this in children, and none have investigated the reporting accuracy of fish intake in children using biomarkers. In a Danish school meal study, 8- to 11-year-old children (n 834) were served the New Nordic Diet (NND) for lunch. The present study examined the accuracy of self-reported intake of signature foods (berries, cabbage, root vegetables, legumes, herbs, potatoes, wild plants, mushrooms, nuts and fish) characterising the NND. Children, assisted by parents, self-reported their diet in a Web-based Dietary Assessment Software for Children during the intervention and control (packed lunch) periods. The reported fish intake by children was compared with their ranking according to fasting whole-blood EPA and DHA concentration and weight percentage using the Spearman correlations and cross-classification. Direct observation of school lunch intake (n 193) was used to score the accuracy of food-reporting as matches, intrusions, omissions and faults. The reporting of all lunch foods had higher percentage of matches compared with the reporting of signature foods in both periods, and the accuracy was higher during the control period compared with the intervention period. Both Spearman's rank correlations and linear mixed models demonstrated positive associations between EPA+DHA and reported fish intake. The direct observations showed that both reported and real intake of signature foods did increase during the intervention period. In conclusion, the self-reported data represented a true increase in the intake of signature foods and can be used to examine dietary intervention effects.
BACKGROUND: Several epidemiologic studies have demonstrated an association between heavy consumption of nonnarcotic analgesics and the occurrence of chronic renal failure, but it is unclear which is the cause and which is the effect METHODS: In a nationwide, population-based, case-control study of early-stage chronic renal failure in Sweden, face-to-face interviews were conducted with 926 patients with newly diagnosed renal failure and 998 control subjects, of whom 918 and 980, respectively, had complete data. We used logistic-regression models to estimate the relative risks of disease-specific types of chronic renal failure associated with the use of various analgesics RESULTS: Aspirin and acetaminophen were used regularly by 37 percent and 25 percent, respectively, of the patients with renal failure and by 19 percent and 12 percent, respectively, of the controls. Regular use of either drug in the absence of the other was associated with an increase by a factor of 2.5 in the risk of chronic renal failure from any cause. The relative risks rose with increasing cumulative lifetime doses, rose more consistently with acetaminophen use than with aspirin use, and were increased for most disease-specific types of chronic renal failure. When we disregarded the recent use of analgesics, which could have occurred in response to antecedents of renal disease, the associations were only slightly attenuated CONCLUSIONS: Our results are consistent with the existence of exacerbating effects of acetaminophen and aspirin on chronic renal failure. However, we cannot rule out the possibility of bias due to the triggering of analgesic consumption by predisposing conditions.
Comment In: N Engl J Med. 2001 Dec 20;345(25):1844-611752364
Comment In: N Engl J Med. 2002 May 16;346(20):1588-9; author reply 1588-912015402
Comment In: N Engl J Med. 2002 May 16;346(20):1588-9; author reply 1588-912017163
To develop algorithm equations that could be used to adjust self-reported height and weight to elicit better estimates of actual BMI.
Linear regression analyses were performed to generate equations that could predict actual height and weight from self-reported data collected through telephone interviews on a representative sample of Canadians aged 18 years or older.
There were systematic biases in self-reported height and weight, leading to an underestimation of BMI. The application of our calibration equations to self-reported data produced closer estimates to actual rates of overweight and obesity.
We advocate the use of our correction equation whenever dealing with self-reported height and weight from telephone surveys to avoid potential distortions in estimating obesity prevalence.
This paper discusses the misclassification that occurs when relying solely on routine register data in family studies of disease clustering. A register study of familial aggregation of schizophrenia is used as an example. The familial aggregation is studied using a regression model for the disease in the child including the disease status of the parents as a risk factor. If all the information is found in the routine registers then the disease status of the parents is only known from the time when the register started and if this information is used unquestioningly the parents who have had the disease before this time are misclassified as disease-free. Two methods are presented to adjust for this misclassification: regression calibration and an EM-type algorithm. These methods are used in the schizophrenia example where the large effect of having a schizophrenic mother hardly shows any signs of bias due to misclassification. The methods are also studied in simulations showing that the misclassification problem increases with the disease frequency.
The aim was to describe quality of care from a patient perspective among adolescents receiving orthodontic treatment and to assess the relationship between quality of care and outcome-related aspects. The research design was cross-sectional. The sample consisted of 151 young people (mean age 17.1 years, SD: 2.2; 53% girls and 47% boys) receiving orthodontic treatment in the Stockholm region in Sweden (response rate 75%). Data were collected using the Quality from the Patient's Perspective questionnaire. The highest quality of care perceptions were noted on items dealing with receiving the best possible orthodontic treatment and being treated with respect. Less favourable perceptions of the quality of care were found regarding the opportunity to participate in the decisions related to the orthodontic treatment. In order to improve the quality of care a more active involvement of these patients in the decision-making process is suggested. The item 'I received the best possible orthodontic treatment' noted the highest subjective importance rating. The youngest participants reported the most favourable scores and the oldest the least. The majority (74%) reported that they were 'completely satisfied' with the result of the orthodontic treatment. However, 52% claimed that they had not followed all of the advice obtained during the treatment period, and 29% indicated some or more hesitation about attending the same dentist for future treatment.
Alcohol policies around the world seek to delay the initiation of drinking. This is partly based on the influential idea that earlier initiation is likely to cause adult alcohol problems. This study synthesises robust evidence for this proposition.
Systematic review of prospective cohort studies in which adolescent measurement of age of first drink in general population studies was separated by at least 3 years from adult alcohol outcomes. EMBASE, Medline, PsychINFO and Social Policy and Practice were searched for eligible studies, alongside standard non-database data collection activities. Data were extracted on included study methods and findings. Risk of bias and confounding was assessed for individual studies and a narrative synthesis of findings was performed.
The main finding was the meagre evidence base available. Only five studies were eligible for inclusion in this review. The existence of effects of age of first drink on adult drinking and related problems were supported, but not at all strongly, in some included studies, and not in others. Rigorous control for confounding markedly attenuates or eliminates any observed effects.
There is no strong evidence that starting drinking earlier leads to adult alcohol problems and more research is needed to address this important question. Policy makers should, therefore, reconsider the justification for delaying initiation as a strategy to address levels of adult alcohol problems in the general population, while also addressing the serious acute harms produced by early drinking.
Cites: Am J Psychiatry. 2000 May;157(5):745-5010784467