Chukiat Viwatwongkasem

Revisiting proportion estimators

Proportion estimators are quite frequently used in many application areas. The conventional proportion estimator (number of events divided by sample size) encounters a number of problems when the data are sparse as will be demonstrated in various settings. The problem of estimating its variance when sample sizes become small is rarely addressed in a satisfying framework. Specifically, we have in mind applications like the weighted risk difference in multicenter trials or stratifying risk ratio estimators (to adjust for potential confounders) in epidemiological studies. It is suggested to estimate p using the parametric family (see PDF for character) and p(1 - p) using (see PDF for character), where (see PDF for character). We investigate the estimation problem of choosing c 0 from various perspectives including minimizing the average mean squared error of (see PDF for character), average bias and average mean squared error of (see PDF for character). The optimal value of c for minimizing the average mean squared error of (see PDF for character) is found to be independent of n and equals c = 1. The optimal value of c for minimizing the average mean squared error of (see PDF for character) is found to be dependent of n with limiting value c = 0.833. This might justifiy to use a near-optimal value of c = 1 in practice which also turns out to be beneficial when constructing confidence intervals of the form (see PDF for character).

Trends in three decades of HIV/AIDS epidemic in Thailand by nonparametric backcalculation method.

Objectives:To reconstruct the past HIV incidence and prevalence in Thailand from 1980 to 2008 and predict the countrys AIDS incidence from 2009 to 2011. Methods:Nonparametric backcalculation was adopted utilizing 100 quarterly observed new AIDS counts excluding pediatric cases. The accuracy of data was enhanced through a series of data adjustments using the weight method to account for several surveillance reporting issues. The mixture of time-dependent distributions allowed the effects of age at seroconversion and antiretroviral therapy to be incorporated simultaneously. Sensitivity analyses were conducted to assess model variations that were subject to major uncertainties. Future AIDS incidence was projected for various predetermined HIV incidence patterns. Results:HIV incidence in Thailand reached its peak in 1992 with approximately 115 000 cases. A steep decline thereafter discontinued in 1997 and was followed by another strike of 42 000 cases in 1999. The second surge, which happened concurrently with the major economic crisis, brought on 60 000 new infections. As of December 2008, more than 1 million individuals had been infected and around 430 000 adults were living with HIV corresponding to a prevalence rate of 1.2%. The incidence rate had become less than 0.1% since 2002. The backcalculated estimates were dominated by postulated median AIDS progression time and adjustments to surveillance data. Conclusion:Our analysis indicated that, thus far, the 1990s was the most severe era of HIV/AIDS epidemic in Thailand with two HIV incidence peaks. A drop in new infections led to a decrease in recent AIDS incidence, and this tendency is likely to remain unchanged until 2011, if not further.

A generalization of Chao's estimator for covariate information.

This note generalizes Chaos estimator of population size for closed capture-recapture studies if covariates are available. Chaos estimator was developed under unobserved heterogeneity in which case it represents a lower bound of the population size. If observed heterogeneity is available in form of covariates we show how this information can be used to reduce the bias of Chaos estimator. The key element in this development is the understanding and placement of Chaos estimator in a truncated Poisson likelihood. It is shown that a truncated Poisson likelihood (with log-link) with all counts truncated besides ones and twos is equivalent to a binomial likelihood (with logit-link). This enables the development of a generalized Chao estimator as the estimated, expected value of the frequency of zero counts under a truncated (all counts truncated except ones and twos) Poisson regression model. If the regression model accounts for the heterogeneity entirely, the generalized Chao estimator is asymptotically unbiased. A simulation study illustrates the potential in gain of bias reduction. Comparisons of the generalized Chao estimator with the homogeneous zero-truncated Poisson regression approach are supplied as well. The method is applied to a surveillance study on the completeness of farm submissions in Great Britain.

A Comparison of Population Size Estimators under the Truncated Count Model With and Without Allowance for Contaminations

The purpose of the study is to estimate the population size under a homogeneous truncated count model and under model contaminations via the Horvitz-Thompson approach on the basis of a count capture-recapture experiment. The proposed estimator is based on a mixture of zero-truncated Poisson distributions. The benefit of using the proposed model is statistical inference of the long-tailed or skewed distributions and the concavity of the likelihood function with strong results available on the nonparametric maximum likelihood estimator (NPMLE). The results of comparisons, for finding the appropriate estimator among McKendricks, Mantel-Haenszels, Zeltermans, Chaos, the maximum likelihood, and the proposed methods in a simulation study, reveal that under model contaminations the proposed estimator provides the best choice according to its smallest bias and smallest mean square error for a situation of sufficiently large population sizes and the further results show that the proposed estimator performs well even for a homogeneous situation. The empirical examples, containing the cholera epidemic in India based on homogeneity and the heroin user data in Bangkok 2002 based on heterogeneity, are fitted with an excellent goodness-of-fit of the models and the confidence interval estimations may also be of considerable interest.

Factors influencing the presence of peripheral arterial disease among Thai patients with type 2 diabetes

Background: Little is known about factors predicting peripheral arterial disease (PAD) development in Thai type 2 diabetes patients. This study aims to identify factors related to PAD in type 2 diabetes and the best predictors for PAD development. Methods and results: A case-control study was conducted in which 405 type 2 diabetes patients were recruited from four tertiary care hospitals in Bangkok, Thailand. Cases were type 2 diabetes patients with PAD who were compared to those without PAD. An ankle-brachial index (ABI) < 0.9 was used for PAD diagnosis. An ABI between 0.91 and 1.30 was used to define those without PAD. Demographic characteristics, the diabetes self-care activities score and chief complaints were measured. Thai type 2 diabetes patients with PAD were older (65.45 ± 12.2 years), had no formal schooling, and were more likely to be a local vendor than those without PAD (p < 0.05). Type 2 diabetes patients with PAD had diabetes longer than those without PAD (p < 0.05). Hypertension, coronary heart disease and chronic kidney disease were significantly more common in type 2 diabetes patients with PAD than those without PAD (p < 0.05). Age > 70 years, having coronary heart disease as a comorbid illness, and having a body mass index of 25–29.9 kg/m2 were predictive for PAD development (all p < 0.05). These three variables explained 12.3% of the variance in the incidence of PAD among type 2 diabetes patients. The demographic and clinical factors were the best predictors for PAD development. Conclusion: Thai type 2 diabetes patients who are elderly, have coronary heart disease as a comorbid condition, or have a normal weight should be considered at risk for PAD development.

PREDICTORS OF INTENTION TO USE HIV TESTING SERVICE AMONG SEXUALLY EXPERIENCED YOUTH IN THAILAND

This study examined the predictors of intention to use HIV counseling and testing (HCT) services among those who had never used HCT services in a sample of 2,536 Thai youth in Bangkok (ages 15-24). Web-based questionnaires included assessments of HIV knowledge, HIV testing attitude, AIDS stigma, and youth-friendly HCT (YFHCT) service expectation. More than 80% of the sexually experienced youth had never used HCT services but among this group 74.06% reported having intentions to do so. The significant predictors consisted of favorable expectations of YFHCT services (p < .001), positive attitude toward HIV testing (p < .005), perceived high risk for HIV infection (p < .01), having multiple sex partners while also using condoms consistently (p < .01), willingness to pay (p < .001), and being informed about HCT and knowing service locations (p < .001). Policy makers, as well as health promotion program developers and researchers can use these findings to increase intention and use of HCT services among at-risk youth.

The Impact of Cigarette Plain Packaging on Health Warning Salience and Perceptions: Implications for Public Health Policy

The study employed a mixed methods design using focus group interviews with 6 student groups and self-administered questionnaires with 1239 students. The participants were nonsmoking, current smokers, and quit-smoking teenagers from secondary schools and colleges. Focus group revealed that although nonsmoking teenagers perceived fear appeals to warning messages, current smokers did not perceive fear appeals to health. Black and white backgrounds of the cigarette package were chosen as the best color for plain packaging. However, most participants suggested various pictorials and a bigger size of pictorial warnings for greater and more effective fear appeal. Odds ratio showed that males had 2.43 times the odds to perceive intention not to smoke. Teenagers who had never smoked and those who had quit smoking had 13.27 and 3.61 times the odds, respectively, to perceive intention not to smoke.

A comparison of risk difference estimators in multi-center studies under baseline-risk heterogeneity

The risk difference is frequently used as a measure of the actual gain in the success rate between two treatments within a center (i.e. hospital). Interest is devoted to combining the risk difference across several centers under homogeneity but allowing for baseline-risk heterogeneity in each of the treatment arms. The purpose is to compare the efficiency of six estimators for the common risk difference. The six estimators consist of the Pooling method ignoring the stratification of centers, several popular sets of different weights, and a new estimator. A simulation study was done to compare bias, variance and mean-square error. The sample sizes in each center varied as 4, 8, 16, 32, 64 and the number of centers as 4, 8, 16, 32, 64. The major result is that the new estimate is an attractive compromise when choosing between the estimators of the set of the center-specific sample size weights and the estimators of the set of the inverse-variance weights. It is not an optimal strategy, but it widely extends to cover heterogeneity cases. For small sample size (n ≤ 8), the Cochran and the Mantel-Haenszel estimators are most efficient because of their smallest mean square errors. Cochran and Mantel-Haenszel estimates are also unbiased and consistent with respect to both sample size and center size. For large sample size (n ≥ 32), Lipsitz et al. and Rothman-Boice estimates whose weights are the inverses of variances are the most appropriate. Lipsitz et al. and Rothman-Boice estimates are considerably biased (even if asymptotically unbiased with respect to the sample size). The Pooling estimate is very close and similar to Cochrans estimate under homogeneity of equal risk difference across centers. We recommend to use Cochran, Mantel-Haenszel, or the Pooling estimators when n ≤ 8, to use Lipsitz et al. and Rothman-Boice estimators when n ≥ 32, and to use the new estimator when strong baseline heterogeneity occurs.

Factors influencing prehospital delay time among patients with peripheral arterial occlusive disease

Background: Only one-third of patients with peripheral arterial occlusive disease (PAOD) seek medical care after perceiving the symptoms of PAOD, and most PAOD patients only visit the physician when they develop ulceration and gangrene. Delay can result in lower extremity amputation and death within three years. The aim of this study was to predict prehospital delay time from sociodemographic characteristics and clinical characteristics, social support, knowledge about PAOD, depression and fear, and treatment-seeking behaviors among patients with PAOD. Method and results: Data were collected in three university hospitals in Bangkok, Thailand. A sample of 212 patients with PAOD who were newly diagnosed or diagnosed within the preceding four months was recruited into the study. Questionnaires and interviewing were used to collect data. Stepwise multiple regression analysis was performed to identify the factors influencing prehospital delay time. Significant determinants of prolonged prehospital delay time were male gender, low monthly income (less than 10,000 Thai baht or 213 Euros), high level of perceived social support, and several treatment seeking behaviors. Depression, high level of fear, and self-pay of medical expenses were associated with short prehospital delay time. Overall, the model explained 41.0% of the variance in prehospital delay time. Conclusion: Clinicians need to develop intervention programs and national campaigns to increase knowledge about PAOD among patients in these high risk groups.

Strategies for Retaining Health-Care Professionals in Rural Areas of India

Background: Retention of skilled health workforce in rural and remote areas is a global problem. While global literature is expanding towards rural retention, however, in the developing countries like India, there is little information available on relevant strategies for implementation of universal healthcare policies and provision of equitable health care distribution. In this review article, the current strategies to retain healthcare professionals are generally be grouped into educational, financial, regulatory and supportive strategies. After that, we document and debate two cadres (i.e. physicians and nurses) involved in primary health care in India and the strategies adopted for improved rural retention in these cadres. We also provide ample evidence to support these strategies and analyze their rationale in augmenting health workforce distribution in India. Lastly, we propose pragmatic ways to deal with future human resource for health reforms. Methods: Literature search was conducted in various electronic databases such as Google Scholar, MEDLINE, PUBMED and EMBASE with relevant key words such as doctors, nurses, health workers, health care professionals, human resources for health, etc. Additional studies were also identified through crossreferences and websites of official agencies. Results: We found that, in educational strategy, reservation for post-graduation seats for medical graduates and in-service physicians (doctors currently working in primary health centres) have been employed in various states of India for rural service. The financial incentive strategy includes incentivizing rural service for both medical doctors and nurses. Under regulatory strategies, better workforce management such as transparent posting mechanisms, shorter recruitment procedures and rotational postings in difficult areas have been employed. Apart from this, other types of health workers such rural medical assistants (RMAs) and alternative medicine (Ayurveda, Yoga, Unani, Sidha and Homeopathy) – collectively known as AYUSH have also been trained and recruited for rural health care practice. Conclusion: In India, rural retention strategies are predominately focused towards physicians. However, state-based evidence and international literature suggest that by providing appropriate financial incentives for rural service to nurses and other cadres such as AYUSH and RMAs is cost-effective and less challenging than the allopathic cadre. Hence, there is an advantage on strengthening rural incentive strategies in these cadres. Further, for these cadres, along with salary, other non-monetary incentives (such as better working conditions, children’s education, good rural living conditions and safety) are needed that can increase rural retention.