Equivalence between Step Selection Functions and Biased Correlated Random Walks for Statistical Inference on Animal Movement.

Animal movement has a fundamental impact on population and community structure and dynamics. Biased correlated random walks (BCRW) and step selection functions (SSF) are commonly used to study movements. Because no studies have contrasted the parameters and the statistical properties of their estimators for models constructed under these two Lagrangian approaches, it remains unclear whether or not they allow for similar inference. First, we used the Weak Law of Large Numbers to demonstrate that the log-likelihood function for estimating the parameters of BCRW models can be approximated by the log-likelihood of SSFs. Second, we illustrated the link between the two approaches by fitting BCRW with maximum likelihood and with SSF to simulated movement data in virtual environments and to the trajectory of bison (Bison bison L.) trails in natural landscapes. Using simulated and empirical data, we found that the parameters of a BCRW estimated directly from maximum likelihood and by fitting an SSF were remarkably similar. Movement analysis is increasingly used as a tool for understanding the influence of landscape properties on animal distribution. In the rapidly developing field of movement ecology, management and conservation biologists must decide which method they should implement to accurately assess the determinants of animal movement. We showed that BCRW and SSF can provide similar insights into the environmental features influencing animal movements. Both techniques have advantages. BCRW has already been extended to allow for multi-state modeling. Unlike BCRW, however, SSF can be estimated using most statistical packages, it can simultaneously evaluate habitat selection and movement biases, and can easily integrate a large number of movement taxes at multiple scales. SSF thus offers a simple, yet effective, statistical technique to identify movement taxis.

Increased HIV prevention program coverage and decline in HIV prevalence among female sex workers in south India.

BACKGROUND: As one way of assessing the impact of Avahan, the India AIDS Initiative of the Bill & Melinda Gates Foundation, we examined the association between HIV prevention program indicators and changes in HIV prevalence among female sex workers (FSWs) between 2005 and 2009. METHODS: We conducted a secondary data analysis from 2 large cross-sectional surveys (2005-2006 and 2008-2009) across 24 districts in south India (n = 11,000 per round). A random-effect multilevel logistic regression analysis was performed using HIV as the outcome, with individual independent variables (from both surveys) at level 1 and district-level FSW-specific program indicators and contextual variables at level 2. Program indicators included their 2006 value, the difference in their values between 2008 and 2006, and the interaction between this difference and study round. RESULTS: HIV prevalence among FSWs decreased from 17.0% to 14.2% (P < 0.001). This decline varied significantly (P = 0.006) across levels of difference in program coverage (% of FSWs contacted by the program in a given year). Odds ratios comparing HIV prevalence between rounds changed with the level of increase in coverage and were statistically significant with coverage increase ≥ quartile (Q) 1: odds ratio, 0.85 at Q1; 0.78 at Q2; 0.66 at Q3; and 0.51 at Q4. CONCLUSIONS: These findings suggest that increased program coverage was associated with declining HIV prevalence among FSWs covered by the Avahan program. The triangulation of our results with those from other approaches used in evaluating Avahan suggests a major impact of this intervention on the HIV epidemic in southern India.

Population-level impact of Avahan in Karnataka state, south India using multilevel statistical modelling techniques.

OBJECTIVE: To assess the population-level impact of "Avahan," the India AIDS Initiative of the Bill & Melinda Gates Foundation, between 2003 and 2008 in Karnataka state, India. DESIGN: Secondary data analysis using all consistent data collection sites from antenatal clinic (ANC) sentinel surveillance data from 2003 to 2008 in Karnataka. METHODS: A multilevel logistic regression model considering individual- and district-level variables was developed to compare time trends in HIV prevalence among young ANC women (younger than 25 years of age) between Avahan (18) and non-Avahan (9) districts. District-level random effects were considered for the intercept and time. The impact was assessed using interaction terms between district type (Avahan vs. non-Avahan) and time. The number of cases averted was estimated, comparing predicted ANC HIV prevalence in the presence versus the absence of Avahan. Data from the National Family Health Survey Round 3 (2006) were used to extrapolate these numbers to the general population. RESULTS: HIV prevalence among young ANC women declined from 1.46% (2003) to 0.83% (2008). The HIV prevalence trend was significantly different between Avahan and non-Avahan districts (P = 0.046). Overall, 87,035 cases of HIV infection were estimated to have been averted in the Karnataka general population because of Avahan during the 2003-2008 period (range under varying assumptions: 55,160-150,784). CONCLUSIONS: Our results suggest that Avahan has had a significant impact on the HIV epidemic in the general population of Karnataka. These results suggest that targeted interventions similar to Avahan should be implemented and scaled up in all concentrated and mixed HIV epidemics.

Variance estimation and confidence intervals for the standardized mortality ratio with application to the assessment of a cancer screening program.

The effect of a cancer screening program can be measured through the standardized mortality ratio (SMR) statistic. The numerator of the SMR is the observed number of deaths from the screened disease among participants in the screening program, whereas the denominator of the SMR is an estimate of the expected number of deaths in these participants under the assumption that the screening program has no effect. In this article, we propose a variance estimator for the denominator of the SMR when this expected number of deaths is estimated with Sasieni's method. We give both a general formula for this variance as well as formulas for specific disease incidence and survival estimators. We show how this new variance estimator can be used to build confidence intervals for the SMR. We investigate the coverage properties of various types of confidence intervals by simulation and find that intervals that make use of the proposed variance estimator perform well. We illustrate the method by applying it to the Québec Breast Cancer Screening program.

Identification of cancer care and protocol characteristics associated with recruitment in breast cancer clinical trials.

PURPOSE: It is estimated that only 5% of patients with cancer participate in a clinical trial. Barriers to participation may relate to available protocols, physicians, and patients, but few data exist on barriers related to cancer care environments and protocol characteristics. METHODS: The primary objective was to identify characteristics of cancer care environments and clinical trial protocols associated with a low recruitment into breast cancer clinical trials. Secondary objectives were to determine yearly recruitment fraction onto clinical trials from 1997 to 2002 in Ontario, Canada, and to compare recruitment fraction among years. Questionnaires were sent to hospitals requesting characteristics of cancer care environments and to cooperative groups/pharmaceutical companies for information on protocols and the number of patients recruited per hospital/year. Poisson regression was used to estimate the recruitment fraction. RESULTS: Questionnaire completion rate varied between 69% and 100%. Recruitment fraction varied between 5.4% and 8.5% according to year. More than 30% of patients were diagnosed in hospitals with no available trials. In multivariate analysis, the following characteristics were associated with recruitment: use of placebo versus not (relative risk [RR] = 0.80; P = .05), nonmetastatic versus metastatic trial (RR = 2.80; P < .01), and for nonmetastatic trials, protocol allowing an interval of 12 weeks or longer versus less than 12 weeks (from diagnosis, surgery, or end of therapy) before enrollment (RR = 1.36; P < .01). CONCLUSION: Allowable interval of 12 weeks or longer to randomly assign patients in clinical trials could help recruitment. In our study, absence of an available clinical trial represented the largest barrier to recruitment.