Moment methods for analysing repeated binary responses have been proposed by Liang and Zeger, and extended by Prentice and Zhao and Prentice. In these estimating equations, models are proposed for the correlation between the repeated binary responses. We extend Liang and Zeger's method to models for the correlation between repeated nominal or ordinal categorical responses; in particular, when the repeated responses are binary, our methods reduce to Liang and Zeger's method. Our method is illustrated with two datasets. One dataset contains repeated observations of self-assessment of arthritis, an ordered variable with three categories, collected during a randomized comparative study of alternative treatments of patients with rheumatoid arthritis. The second dataset is a longitudinal study of the health effects of air pollution, in which the repeated ordered multinomial response is the wheezing status (no wheeze, wheeze with cold, wheeze apart from cold) of a child at ages 9, 10, 11 and 12 years.

No abstract

An estimate of a parameter vector beta is often obtained by setting a "score" vector equal to zero and solving for beta. Estimating equations of this type include maximum likelihood, quasi-likelihood (McCullagh, 1983, Annals of Statistics 11, 59-67), and generalized estimating equations (Liang and Zeger, 1986, Biometrika 73, 13-22). White (1982, Econometrica 50, 1-26) proposed a variance estimator for beta that is robust to model misspecification. We show that a "one-step" jackknife estimator of variance is asymptotically equivalent to the variance estimator proposed by White. The one-step variance estimator may be preferred when the appropriate computer packages are not available to compute White's estimator directly. This jackknife estimator is very useful in our example with clustered survival data.

BackgroundHealth information exchange (HIE) is frequently cited as an important objective of health information technology investment because of its potential to improve quality, reduce cost, and increase patient satisfaction. In this paper we examine the status and practices of HIE in six countries, drawn from a range of higher and lower income regions.MethodsFor each of the countries represented – China, England, India, Scotland, Switzerland, and the United States – we describe the state of current practice of HIE with reference to two scenarios: transfer of care and referral. For each country we discuss national objectives, barriers and plans for further advancing clinical information exchange.ResultsThe countries vary widely in levels of adoption of EHRs, availability of health information in electronic form suitable for HIE, and in the information technology infrastructure to be used for transmission. Common themes emerged, however, including an expectation that information will be exchanged rather than gathered anew, the need for incentives to promote information exchange, and concerns about data security and patient confidentiality.ConclusionsAlthough the ability to transfer health information to where it is most needed is nearly always mentioned as an advantage of HIE adoption, there are wide differences in the degree to which this has been achieved to support the scenarios used in this study. Nevertheless, these differences indicate varying stages of progress along a comparable pathway, with similar barriers being identified in the countries described. In some cases, these have been partially surmounted while elsewhere work is needed. We reflect on contextual factors influencing the status and direction of HIE efforts in different global regions and their implications for progress.

BackgroundSeasonal associations of cardiovascular mortality have been noted in most populations of European origin years ago, but are not well evaluated in Asian populations recently.MethodsUtilizing the electronic Hospitalization Summary Reports (HSRs) from 32 top-ranked hospitals in Beijing, China, we evaluated the association between winter season and the risk of cardiovascular death among hospitalized individuals. General additive models and logistic regression models were adjusted for confounding factors.ResultsOlder patients who were admitted to the hospital in the winter months (January, February, November and December) had a death risk that was increased by approximately 30% to 50% (P < 0.01) over those who were admitted in May. However, younger patients did not seem to experience the same seasonal variations in death risk. The excess winter deaths among older patients were associated with ischemic heart disease (RR = 1.22; 95% CI 1.13 to 1.31), pulmonary heart disease (RR = 1.42; 95% CI 1.10 to 1.83), cardiac arrhythmias (RR = 1.67; 95% CI 1.36 to 2.05), heart failure (RR = 1.30; 95% CI 1.09 to 1.54), ischemic stroke (RR = 1.30; 95% CI 1.17 to 1.43), and other cerebrovascular diseases (RR = 1.78; 95% CI 1.40 to 2.25). The risks of mortality were higher in winter months than in the month of May, regardless of the presence or absence of respiratory disease.ConclusionsWinter season was associated with a substantially increased risk of cardiovascular death among older Chinese cardiovascular inpatients.