Complementary Log-log Regression Research Articles

Causal inference did not detect any effect of jump load on knee complaints in elite men's volleyball.

The aim was to determine how jump load affects knee complaints in elite men's volleyball. We collected data from four men's premier league volleyball teams through three seasons in a prospective cohort study (65 players, 102 player-seasons). Vert inertial measurement devices captured the jump load (jump frequency and jump height) from 21 088 daily player sessions, and knee complaints were reported in 3568 weekly OSTRC-O questionnaires. Mixed complementary log-log regression models described the probability of (i) experiencing symptoms if players were currently asymptomatic, (ii) worsening symptoms if players had symptoms, and (iii) recovery from knee complaints. Based on our causal assumptions, weekly jump load was modeled as the independent variable, adjusted for age (years), weight (kg), position on volleyball team, and past jump load. No certain evidence of an association was found between weekly jump load and probability of (i) knee complaints (p from 0.10 to 0.32 for three restricted cubic splines of load), (ii) worsening symptoms if the player already had symptoms (p from 0.11 to 0.97), (iii) recovery (p from 0.36 to 0.63). The probability of knee complaints was highest for above-average weekly jump load (~1.2% for an outside hitter with mean age and height) compared with low loads (~1%) and very high loads (→ ~ 0%). The association between jump load and knee complaints risk remains unclear. Small differences in risk across observed jump load levels were observed. It would likely require substantially increased sample sizes to detect this association with certainty.

Minimum acceptable diet and its predictors among children aged 6-23 months in Ethiopia. A multilevel cloglog regression analysis.

Despite significant progress made previously and the recognized health benefits of optimal feeding practices, ensuring a minimum acceptable diet in developing countries like Ethiopia remains a formidable challenge. Additionally, there is a scarcity of data in this area. Therefore, our study aims to identify predictors of a minimum acceptable diet using a powerful tool called complementary log-log regression analysis. Thus, it contributes to accelerating the pathway of ending child undernutrition thereby promoting optimal child health. A multilevel analysis was conducted among a weighted sample of 1427 children aged 6-23 months using the 2019 Ethiopian Demographic Health Survey (EDHS). The EDHS sample was stratified and selected in two stages. A minimum acceptable diet is defined as a composite of children fed with both minimum dietary diversity and minimum meal frequency. Data extraction took place between August 1 and 30, 2023. We used STATA software version 17 for data analysis. A complementary log-log regression model was fitted to identify significant predictors of the minimum acceptable diet. A p-value of ≤0.05 was used to declare statistically significant predictors. Only 10.44% (95CI: 8.90-12.15) of the children meet the minimum acceptable diet. Child aged (18-23 month) (AOR = 1.78, 95CI:1.14-2.78)], mother's educational level (secondary and above education) (AOR = 279,95CI: 1.51-5.15), number of children three and above [(AOR = 0.78,95CI: 0.53-0.94], institutional delivery [AOR = 1.77,95CI: 1.11-3.11], having postnatal-check-up [AOR = 2.33,95CI: 1.59-3.41] and high community poverty level (AOR = 0.49,95CI: 0.29-0.85) were found to be predictors of minimum acceptable diet. In Ethiopia, only one in ten children achieve a minimum acceptable diet. Which is lower than the global report findings (16%). Enhancing maternal education programs and promoting family planning strategies to reduce household size are essential. Besides, encouraging institutional deliveries and postnatal check-ups are also recommended. It is necessary to implement targeted interventions for poverty reduction in communities to ensure that families can afford nutritious diets for their children.

Determinants of late antenatal care attendance among high parity women in Uganda: analysis of the 2016 Uganda demographic and health survey

BackgroundTimely and adequate Antenatal Care (ANC) effectively prevents adverse pregnancy outcomes and is crucial for decreasing maternal and neonatal mortality. High-parity women (5 + children) are at higher risk of maternal mortality. Limited information on the late timing of ANC among this risky group continues to hamper Uganda’s efforts to reduce maternal mortality ratios and improve infant and child survival. This study aimed to determine factors associated with attendance of the first ANC after 12 weeks of gestation among high-parity women in Uganda.MethodsThis study was based on nationally representative data from the 2016 Uganda Demographic and Health Survey. The study sample comprised 5,266 women (aged 15–49) with five or more children. A complementary log-log regression model was used to identify factors associated with late ANC attendance among high-parity women in Uganda.ResultsOur findings showed that 73% of high parity women delayed seeking their first ANC visit. Late ANC attendance among high-parity women was associated with distance to the health facility, living with a partner, partner’s education, delivery in a health facility, and Desire for more children. Women who did not find the distance to the health facility when going for medical help to be a big problem had increased odds of attending ANC late compared to women who found distance a big problem (AOR = 1.113, CI: 1.004–1.234), women not living with partners (AOR = 1.196, 95% CI = 1.045–1.370) having had last delivery in a health facility (AOR = 0.812, 95% CI = 0.709–0.931), and women who desired to have another child (AOR = 0.887, 95% CI = 0.793–0.993) had increased odds compared to their counterparts.ConclusionsTo increase mothers’ timely attendance and improve maternal survival among high-parity women in Uganda, programs could promote and strengthen health facility delivery and integrate family planning with other services such as ANC and postnatal care education to enable women to seek antenatal care within the recommended first trimester. This study calls for increased support for programs for education, sensitization, and advocacy for health facility-based deliveries. This could be done through strengthened support for VHT and community engagement activities.

Prevalence and associations between occupational diseases, emotional exhaustion, and dust mask anxiety among coal miners in Northern China.

This study examined the prevalence and associations between occupational-related diseases, emotional exhaustion, and dust mask anxiety among coal miners in Northern China. The study sampled 506 miners from two coal mines in Northern China. Descriptive (frequency, percentages, and chi-square tests) and inferential (complementary log-log logistic regression) analyses were carried out to assess prevalence and associations. The prevalence of coal mining-related pulmonary diseases was 94%, with 40% and 37% reporting pneumoconiosis and bronchitis respectively. Emphysema and tuberculosis were less prevalent with a total prevalence of 5% while 12% of the miners reported two, three, or four of these diseases. For cardiovascular diseases, 66% of the miners reported 3H (hypertension, hyperlipemia, and hyperglycemia) while heart disease and cerebral infarction were reported by 6% and 2% of the miners respectively. Also, 5% and 2% of the miners reported two or all three of 3H, heart condition, and cerebral infarction. The overall prevalence of cardiovascular diseases was 81% while 82% and 63% of the miners reported experiencing some frequency of emotional exhaustion and dust mask anxiety respectively. The study also identified associations between mining-related diseases, emotional exhaustion, dust mask anxiety, and other demographic, personal habits and work-related factors. Miners with pulmonary diseases were more likely to experience emotional exhaustion and dust mask anxiety. Contrariwise, miners with cardiovascular diseases were less likely to experience dust mask anxiety. Factors such as age, engaging in physical exercises, rhinitis, source of pressure and shift system were also associated with emotional exhaustion and dust mask anxiety. These findings highlight the prevalence of occupational diseases, emotional exhaustion, and dust mask anxiety among coal miners in China. The study emphasizes the need for interventions to address health risks, improve work conditions, and support miners' well-being in the coal mining industry.

Sinus inflammation and chronic rhinosinusitis are associated with a diagnosis of new onset asthma in the following year.

Chronic rhinosinusitis (CRS) and asthma commonly co-occur. No studies have leveraged large samples needed to formally address whether preexisting CRS is associated with new onset asthma over time. We evaluated whether prevalent CRS [identified in two ways: validated text algorithm applied to sinus computerized tomography (CT) scan or two diagnoses] was associated with new onset adult asthma in the following year. We used electronic health record data from Geisinger from 2008 to 2019. For each year we removed persons with any evidence of asthma through the end of the year, then identified those with new diagnosis of asthma in the following year. Complementary log-log regression was used to adjust for confounding variables (e.g., sociodemographic, contact with the health system, comorbidities), and hazard ratios (HRs) and 95% confidence intervals (CI) were calculated. A total of 35,441 persons were diagnosed with new onset asthma and were compared to 890,956 persons who did not develop asthma. Persons with new onset asthma tended to be female (69.6%) and younger (mean [SD] age 45.9 [17.0] years). Both CRS definitions were associated (HR, 95% CI) with new onset asthma, with 2.21 (1.93, 2.54) and 1.48 (1.38, 1.59) for CRS based on sinus CT scan and two diagnoses, respectively. New onset asthma was uncommonly observed in persons with a history of sinus surgery. Prevalent CRS identified with two complementary approaches was associated with a diagnosis of new onset asthma in the following year. The findings may have clinical implications for the prevention of asthma.

Abstract P303: Time-Varying Healthy Lifestyle Profile and All-Cause Mortality Among the Oldest-Old in China

Background: Association between an individual modifiable lifestyle with mortality is well established. However, influences of combined lifestyles on mortality risks remain inconclusive among adults aged 80 years or older (“oldest-old”). We investigated the association of time-varying healthy lifestyle profile (HLP) with mortality among Chinese oldest-old adults. Methods: We pooled 36,021 oldest-old adults from first seven waves of the Chinese Longitudinal Healthy Longevity Survey (1998-2014), with follow-up to latest wave (2018) (range 0-21.04 years). Lifestyle information was collected at each wave. Nonsmoking, no regular alcohol intake, regular exercise, and simplified healthy eating index of 11-18 were defined as healthy lifestyle. We assigned a score of 1 for healthy and 0 for nonhealthy lifestyle and summed the four scores to a HLP index ranging 0-5 (representing least to most healthy) to reflect joint lifestyle measurements. We used both a continuous time-to-event data set up with study enrollment as time origin (SE model) and a discrete time-to-event data set up with age 80 as time origin. We used Cox proportional hazard model with time-varying covariates and complementary log-log regression to examine hazard ratios (HRs) of mortality, adjusting for demographics, self-reported diseases, activities of daily living, and cognitive function. Results: During 110897.40 person-years of follow-up, 25,511 deaths were recorded. Median death age was 5 years older in most versus least healthy group. In SE model with least healthy group as reference, HRs (95% CIs) for mortality were 0.90 (0.86-0.95), 0.83 (0.79-0.87), 0.74 (0.70-0.78), 0.54 (0.49-0.58), respectively for HLP of 1-4 groups (p-trend<0.001) ( Table ). Dose-response relationships were consistent for HRs of mortality over age since 80. Conclusion: Healthier lifestyle profiles were associated with significant reductions in all-cause mortality risks among Chinese oldest-old, suggesting the importance of life-long adherence to HLP into advanced old age.

Association between health and safety perceptions of COVID-19 vaccine and its uptake in Ghana.

Attitudes towards vaccines have affected COVID-19 vaccination programs in many countries. This study sought to evaluate the effects of general perceptions on the safety and health concerns and the confidence in COVID-19 vaccines on its uptake in Ghana. A cross-sectional online survey was conducted between January and March 2021. The outcome variables for this study were "Taking mandatory COVID-19 vaccine" and "Taking voluntary COVID-19 vaccine". The data were subjected to both descriptive (frequency, percentages, and chi-square tests) and inferential (complementary log-log logistic regression) analyses. Out of 620 Ghanians who participated in the survey, about 80% of the participants believed that vaccines were good for one's health and 73% had confidence on COVID-19 vaccine safety; although 81% of the respondents were particularly concerned about the source of the vaccine. 79% and 71% of respondents indicated their willingness for mandatory and voluntary COVID-19 vaccination, respectively. In all operationalized regression models, Ghanaians who believed that vaccines are healthy (OR=1.998, Cl=1.345-2.968; OR=1.652, Cl=1.050-2.601) and those who had confidence in a COVID-19 vaccine safety (OR=4.405, Cl=3.136-6.188; OR=8.340, Cl=5.471-12.713) were more likely to take a mandatory or voluntary COVID-19 vaccine compared to those who thought and believed otherwise. Individual preferences and/or intentions towards COVID-19 vaccine uptake and uptake route (i.e., mandatory, voluntary) were influenced by multifaceted determinants: biosocial (age, marital status, education), socio-cultural (religion, source of vaccine as a concern), and location (geographical zone) factors. To consolidate and possibly increase vaccine uptake in response to the COVID-19 pandemic in Ghana, health education and promotion programs should aim at creating awareness on the benefits of vaccine uptake while addressing the health and safety concerns on the potential side effects through evidence-based community messaging from credible sources. It is important to show specific commitment to transparency and reliable information to build public trust by decision-makers.

Role of maternal and child health services on the uptake of contraceptive use in India: A reproductive calendar approach.

BackgroundAccording to the latest round of National Family Health Survey—4 (NFHS (2015–16)) maternal and child health care (MCH) services improved drastically compared to NFHS-3. Previous studies have established that the uptake of MCH services increases the likelihood of early adoption of contraceptives among women. So, our study aims to examine if the early initiation of contraceptive has proportionately improved with the recent increase in MCH services.MethodsThis study used the reproductive calendar of NFHS-4, 2015–16, to evaluate contraceptive initiation within 12 months after the last birth among 1,36,962 currently married women in India. A complementary log-log regression model was created to examine the link between the time of initiation of contraception and MCH care at the national level.ResultsIt was found that only a quarter of women within 12 months from last birth have adopted the modern contraceptive method. Among those majority of the females adopted sterilization mostly at the time of birth. The multivariable model identified, that the period of initiation of contraceptive depends on the gender composition of children and access to MCH services. It was found that the odds of early initiation of contraceptive use was higher when a women have only son (AOR = 1.15,95% CI– 1.22, 1.18) compared to women with only daughter. Also, it was found that women who have availed MCH services were more likely to adopt contraceptives earlier.ConclusionThe number of women availing MCH services has increased in India, but it did not result in a proportional increase in initiation of contraception after childbirth. Facilitating family planning services alongside MCH services will be beneficial in low-resource settings. It is a golden opportunity to educate and encourage women for early adoption of contraceptive.

Persistence on oral pre-exposure prophylaxis (PrEP) among female sex workers in eThekwini, South Africa, 2016-2020.

Despite the established efficacy of PrEP to prevent HIV and the advantages of a user-controlled method, PrEP uptake and persistence by women in both trials and demonstration projects has been suboptimal. We utilized real-world data from an HIV service provider to describe persistence on oral PrEP among female sex workers (FSW) in eThekwini, South Africa. We examined time from PrEP initiation to discontinuation among all FSW initiating PrEP at TB HIV Care in eThekwini between 2016-2020. We used a discrete time-to-event data setup and stacked cumulative incidence function plots, displaying the competing risks of 1) not returning for PrEP, 2) client discontinuation, and 3) provider discontinuation. We calculated hazard ratios using complementary log-log regression and sub-hazard ratios using competing risks regression. The number of initiations increased each year from 155 (9.3%, n = 155/1659) in 2016 to 1224 (27.5%, n = 1224/4446) in 2020. Persistence 1-month after initiation was 53% (95% CI: 51%-55%). Younger women were more likely to discontinue PrEP by not returning compared with those 25 years and older. Risk of discontinuation through non-return declined for those initiating in later years. Despite the COVID-19 pandemic, a greater number of initiations and sustained persistence were observed in 2020. Low levels of PrEP persistence were observed, consistent with data among underserved women elsewhere. Encouragingly, the proportion of women persisting increased over time, even as the number of women newly initiating PrEP and staff workload increased. Further research is needed to understand which implementation strategies the program may have enacted to facilitate these improvements and what further changes may be necessary.

Characterizing and mapping the spatial variability of HIV risk among adolescent girls and young women: A cross-county analysis of population-based surveys in Eswatini, Haiti, and Mozambique.

BackgroundTo stem the HIV epidemic among adolescent girls and young women (AGYW), prevention programs must target services towards those most at risk for HIV. This paper investigates approaches to estimate HIV risk and map the spatial heterogeneity of at-risk populations in three countries: Eswatini, Haiti and Mozambique.MethodsWe analyzed HIV biomarker and risk factor data from recent population-based household surveys. We characterized risk using three approaches: complementary log-log regression, latent class analysis (LCA), and presence of at least one risk factor. We calculated the proportion and 95 percent confidence intervals of HIV-negative AGYW at risk across the three methods and employed Chi-square tests to investigate associations between risk classification and HIV status. Using geolocated survey data at enumeration clusters and high-resolution satellite imagery, we applied algorithms to predict the number and proportion of at-risk AGYW at hyperlocal levels.ResultsThe any-risk approach yielded the highest proportion of at-risk and HIV-negative AGYW across five-year age bands: 26%-49% in Eswatini, 52%-67% in Haiti, and 32%-84% in Mozambique. Using LCA, between 8%-16% of AGYW in Eswatini, 37%-62% in Haiti, and 56%-80% in Mozambique belonged to a high vulnerability profile. In Haiti and Mozambique, the regression-based profile yielded the lowest estimate of at-risk AGYW. In general, AGYW characterized as “at risk” across the three methods had significantly higher odds of HIV infection. Hyperlocal maps indicated high levels of spatial heterogeneity in HIV risk prevalence and population density of at-risk AGYW within countries.ConclusionCharacterizing risk among AGYW can help HIV prevention programs better understand the differential effect of multiple risk factors, facilitate early identification of high-risk AGYW, and design tailored interventions. Hyperlocal mapping of these at-risk populations can help program planners target prevention interventions to geographic areas with populations at greatest risk for HIV to achieve maximal impact on HIV incidence reduction.

Knee symptom but not radiographic knee osteoarthritis increases the risk of falls and fractures: results from the Osteoarthritis Initiative

To describe the effect of knee symptoms and radiographic osteoarthritis (ROA) on the risk of falls, recurrent falls, and fractures. Participants from the Osteoarthritis Initiative were classified as having 'no', 'unilateral' or 'bilateral' knee symptoms (≥19 on a 0-96 Western Ontario and McMaster Universities Osteoarthritis Index) and ROA (Kellgren-Lawrence grade ≥2) for each visit. Self-reported falls and fractures in the past 12 months were extracted at baseline and follow-up visits until month 96. Recurrent falls were defined as having ≥2 falls in the past 12 months. Hazard ratios (HR) with 95% confidence intervals (CI) were estimated using mixed-effects complementary log-log regression. Of 4465 participants, 3145 (70%), 1681 (38%), and 806 (18%) experienced at least one fall, recurrent fall, and fracture, respectively, over 96 months. Compared to participants without symptomatic knee, unilateral and bilateral knee symptoms were associated with a 17% increased risk of falls and a 36-46% increased risk of recurrent falls, and bilateral knee symptoms increased the risk of fractures (HR 1.45, 95%CI 1.17 to 1.81). Compared to participants with no ROA in either knee, bilateral ROA was associated with a reduced risk of falls (HR 0.87, 95%CI 0.77 to 0.99) and fractures (HR 0.78, 95%CI 0.64 to 0.96). No statistically significant interactions between knee symptoms and ROA were observed. This large population-based study showed that knee symptoms but not ROA increased the risk of falls, recurrent falls, and fractures, and that adults with bilateral ROA may have a lower risk of falls and fractures.

Drug Interactions and Juvenile Delinquency in Mexico

ABSTRACT We analyze the interaction between Marijuana and Alcohol as correlates of property crimes among Mexican adolescents. Data were drawn from two waves of a Mexican national survey of public high school students in Mexico. A complementary log-log multilevel regression was fitted separately for each survey year, controlling for neighborhood, school, family, and demographic characteristics. We found that Marijuana users were more likely to self-report the commission of a property crime than alcohol users. Other contextual variables such as perceived neighborhood unsafety, drug dealing in school, and disrupted family contexts also increased the likelihood of juvenile delinquency. The absence of evidence of an interaction between the two drugs suggests that their combination, even when they may have even co-occurred, made no significant difference in predicting their involvement in property crimes. This study shows that drug use and juvenile delinquency prevention programs cannot be implemented without consideration of school-context characteristics.

Are structural abnormalities on knee MRI associated with osteophyte development? Data from the Osteoarthritis Initiative

To assess which structural abnormalities on knee MRI are associated with development of osteophytes in middle-aged subjects without radiographic knee osteoarthritis. We included subjects from the Osteoarthritis Initiative, aged 40-55 years, Kellgren & Lawrence grade 0 in both knees, and knee MRIs from both knees available at baseline, 24, 48 and 72 months. Structural exposures on MRI assessed using MOAKS included cartilage damage, bone marrow lesions, meniscal tear, meniscal extrusion, and Hoffa/effusion synovitis. We assessed whether each structural exposure was associated with the development of osteophytes on MRI in the medial and lateral tibiofemoral, and patellofemoral compartment. We estimated hazard ratios (HR) including 95% confidence intervals (CI) for osteophyte development using a mixed complementary log-log regression model adjusted for age, sex, and body mass index. We included 680 knees from 340 subjects with a mean (SD) age of 50 years (3.0), and 51% men. In the medial tibiofemoral compartment, the absolute risk of osteophyte development in the first 24-month period was 4% in knees without, and 15% in knees with medial meniscal tear. Corresponding adjusted HR was 6.6 (95%CI=3.4-12.9). In the lateral tibiofemoral compartment, the adjusted HR for developing osteophytes having a lateral meniscal tear was 3.3 (95%CI=1.3-8.4). In the patellofemoral compartment, patellofemoral cartilage damage was most clearly associated with developing osteophytes (HR=2.6, 95%CI=1.8-3.7). Meniscal tear seem to be the strongest structural risk factor for the development of tibiofemoral osteophytes, and patellofemoral cartilage damage for the development of patellofemoral osteophytes, respectively. Local biomechanical factors are important in early osteophyte development.

Abstract 61: Factors Related to Human Papillomavirus Vaccine Uptake and Acceptability in Manizales, Colombia, 2017-2019: A Population-Based Study of the 2003, 2004 and 2005 Birth Cohorts.

Abstract Purpose: The first 2 rounds (2012-2013) of the Colombian school-based HPV vaccination program attained 80% coverage of nine-year-old girls. After a psychogenic event in 2014 in a town of the Caribbean region, vaccination fell nationwide to 20% in 2016. We identified factors associated to decreasing coverage of the HPV vaccination program in Manizales, Colombia. Methods: Structured questionnaires based on the HPV health belief model (HBM) were used in a population-based survey (September 2017-February 2019) to privately interview girls from birth cohorts of 2003, 2004 and 2005 and their parents. Vaccination coverage (at least one dose) was defined by the Expanded Program of Immunization Registry, HPV 6/11/16/18 serology and parents and girls self-report. Multivariable adjusted Prevalence Ratios (aPR) and 95% confidence intervals were estimated using complementary log-log regression to identify sociodemographic factors associated to coverage and changes in HBM constructs scores associated to acceptability of the HPV vaccine. Results: 1.287 of 1.299 eligible girls and their parents completed the survey, 99% were school-registered and 97% urban area residents. The vaccine coverage was 93.4% (394/422), 84% (356/424), and 63% (279/441) for the 2003, 2004, and 2005 birth cohorts, respectively. High educational level in parents was associated with decreased HPV vaccine coverage (aPR 0.80, 95%CI 0.66-0.95). The perceived benefits increased (aPR 1.22, 95%CI 1.18-1.27 for parents; aPR 1.25, 95%CI 1.21-1.29 for girls) meanwhile, the perceived barriers (negative recommendation by family/friends, news about vaccine adverse events and health system and school barriers) decreased (aPR 0.73 95%CI 0.69-0.77 for parents; aPR 0.84 95%CI 0.79-0.88 for girls) the acceptability of HPV vaccine. Birth year (2005) and high socioeconomic status were also associated with decrease acceptability. Conclusion: Perceived barriers were associated to decreased acceptability and coverage of vaccination in the first 3 rounds of the national HPV vaccination program in Manizales, Colombia. Citation Format: Karen Cárdenas-Garzón, Maria Cecilia Agudelo, Olga Lucia Tovar-Aguirre, Sandra Franco, Kelly Paola Valverde, Nelson Arias, Difariney Gonzalez, Nilton E. Montoya, Maria Patricia Arbelaez, Isabel Cristina Garces-Palacio, Karly Louie, Simon Beddows, Gloria Sanchez. Factors Related to Human Papillomavirus Vaccine Uptake and Acceptability in Manizales, Colombia, 2017-2019: A Population-Based Study of the 2003, 2004 and 2005 Birth Cohorts. [abstract]. In: Proceedings of the 9th Annual Symposium on Global Cancer Research; Global Cancer Research and Control: Looking Back and Charting a Path Forward; 2021 Mar 10-11. Philadelphia (PA): AACR; Cancer Epidemiol Biomarkers Prev 2021;30(7 Suppl):Abstract nr 61.

Pre-Stroke Depression in Ghana and Nigeria: Prevalence, Predictors and Association With Poststroke Depression.

Depression is a risk factor for stroke. There is a knowledge gap on the predictors of prestroke depression in stroke survivors living in low- and middle-income countries (LMICs). We estimated prevalence and predictors of prestroke depression, as well as its association with poststroke depression (PSD) in the largest study of stroke in Africa. We evaluated information collected as part of the Stroke Investigative Research and Education Network (SIREN) study, a multicentre, case-control study conducted at 15 sites in Ghana and Nigeria. Prestroke depression status was ascertained in stroke survivors using a validated self-report tool, while PSD was assessed using a stroke specific screening tool for depression ("HRQOLISP-E"). Independent associations were investigated using complementary log-log regression and binary logit models. Among 1,977 participants, prestroke depression was found in 141 (7.1%). In multivariate analyses, prestroke depression was significantly associated with tachycardia (OR = 2.22, 95% CI = 1.37-3.56) and low consumption of green leafy vegetables (OR = 1.91, 95% CI = 1.12-3.24). Forty-one (29.1%) of the prestroke depression sub-sample developed PSD. However, prestroke depression was not significantly associated with PSD. The findings should energize before-the-stroke identification and prioritization of limited treatment resources in LMICs to persons with depression who have multiple, additional, risks of stroke.

Determinants of Households’ Access to Improved Drinking Water Sources: A Secondary Analysis of Eswatini 2010 and 2014 Multiple Indicator Cluster Surveys

Worldwide, millions of people still die from diseases associated with inadequate water supply, sanitation, and hygiene, despite the fact that the United Nations recognized access to clean drinking water and sanitation as a human right nearly a decade ago. The objective of this study was to describe the determinants of access to improved drinking water sources in Eswatini in 2010 and 2014. Using the Eswatini Multiple Indicator Cluster Surveys (EMICSs), data for 4,819 households in 2010 and 4,843 in 2014 were analyzed. Bivariate and multivariate complementary log-log regression analyses were conducted to identify the determinants of households’ access to improved drinking water sources. The study found that households’ access to improved drinking water sources significantly improved from 73.1% in 2010 to 77.7% in 2014 (p<0.0001). In 2010, households whose heads were aged 35–54 and 55 years had lower odds of having access to improved drinking water sources than those with younger ones. In 2014, female-headed households had lower odds, while, in 2010, sex of the household head was not associated with access to improved drinking water sources. In both years, an increase in the number of household members was negatively associated with access to improved drinking water sources compared to those with fewer members. In both years, the odds of access to improved drinking water sources increased with an increase in the wealth index of the household, and households located in urban areas had higher odds of access to improved drinking water sources compared to those in rural settings. In both years, households from the Shiselweni and Lubombo regions had lower odds of access to improved drinking water sources. The government and its partners should continue to upscale efforts aimed at increasing access to improved drinking water, especially in rural areas, to reduce the disparity that exists between urban and rural households.

On the Statistical Testing Methods for Single Laboratory Validation of Qualitative Microbiological Assays with an Unpaired Design.

There exists several statistical methods for detecting a difference of detection rates between alternative and reference qualitative microbiological assays in a single laboratory validation study with an unpaired design. We compared performance of eight methods including Fisher's exact test, unequal variance two-sample t-test, Wilcoxon rank-sum test, z-test, and methods based on Wilson confidence intervals, complementary log-log regression, Firth's logistic regression, and ordinary logistic regression. We first compared the minimum detectable difference in the proportion of detections between the alternative and reference methods among these statistical methods for a varied number of test portions. We then compared power and size of test of these methods using simulated data. Firth's logistic regression and the unequal variance two-sample t-test had the lowest minimum detectable difference and highest power. None of these statistical methods had an estimated size of test always within a 95% confidence interval of the nominal value 0.05 with small numbers of test portions (n = 12, 20, 30). Fisher's exact test, the Wilcoxon rank-sum test, and the z-test were conservative even with a moderately large number of test portions (n = 40), while Firth's logistic regression and the unequal variance two-sample t-test had a size of test closer to 0.05 than other methods. Firth's logistic regression and the unequal variance two-sample t-test are better choices than other competing methods. We recommend the unequal variance two-sample t-test over Firth's logistic regression because the unequal variance two-sample t-test is better known and easier to use. We provide an example using real data.

To Move or not to Move: Community Members’ Reaction to Surface Mining Activities in the Upper West Region of Ghana

Despite the impact of mining-induced environmental change on community livability, we know little about how disparities in knowledge of health risks associated with mining influence residents’ response, especially in an already environmentally stressed context. Guided by theoretical insights from solastalgia, we examined residents’ decision to relocate due to increasing gold mining activities in the fragile Northern Savannah Ecological Zone of Ghana. Fitting complementary log-log regression models to cross-sectional data from the Upper West Region (UWR) of Ghana, we found that residents with limited knowledge of potential health impacts of mining and those who believe mining activities were not meeting environmental standards were more likely to consider relocating. Given the centrality of land in community health and wellbeing in the UWR, Ghana’s mining guidelines should promote local participation in the regulation of mining activities and guarantee the rights of indigenes to livable native lands.

Cell Phone Counseling Improves Retention of Mothers With HIV Infection in Care and Infant HIV Testing in Kisumu, Kenya: A Randomized Controlled Study.

We evaluated the effectiveness of a cell phone counseling intervention to promote retention in care and HIV testing of infants among women with HIV accessing prevention of mother-to-child services in Kisumu, Kenya. Between May 2013 and September 2015, we recruited 404 pregnant women with HIV who were between 14 and 36 weeks of gestation and randomly assigned them to the intervention (n=207) or control arm (n=197). Retention was assessed at delivery and at 6 and 14 weeks postpartum. We also measured uptake of infant HIV testing. The intervention comprised a fixed protocol of counselor-delivered phone calls to provide one-to-one need-based support. The number of calls made varied depending on when participants presented for antenatal care services; the maximum number was 42. The control group received routine care. We evaluated retention at 3 time points using the complementary log-log regression model taking into account factors associated with retention and loss to follow-up time. We calculated the incidence rate for HIV transmission among infants and used binary logistic regression to identify predictors of HIV infection among infants. Participants attended on average 63% of the required number of counseling calls during the study period. Retention was higher in the intervention arm than the control arm at delivery (95.2% vs. 77.7%, respectively); 6 weeks postpartum (93.9% vs. 72.9%, respectively); and 14 weeks postpartum (83.3% vs. 66.5%, respectively) (P<.001). The counseling intervention (hazard ratio [HR]=0.29; 95% confidence interval [CI]=0.12, 0.69) and positive health perceptions (HR=0.99; 95% CI=0.98, 1.00) were associated with lower hazards of being lost to follow-up. HIV testing of infants was higher in the intervention than control arm (93% vs. 68%, respectively; P<.001). In total, 9 of 308 (2.9%) infants tested positive for HIV infection (incidence rate=0.39 infections/100 infant-weeks). Medication Possession Ratio (MPR) >90%, used to assess adherence to ART, was associated with lower odds of a positive HIV test among infants (adjusted odds ratio=0.20; 95% CI=0.04, 0.99). Attendance at antenatal and postnatal care visits was higher among participants in the intervention arm than the control arm. The one-on-one tailored counseling delivered via cell phone was effective in retaining mothers with HIV infection in care and promoting uptake of infant HIV testing and antenatal and postnatal care services. Phone counseling offers a practical approach to reach and retain pregnant women with HIV infection and postpartum mothers in care, but greater emphasis on collection of medications and adherence is required.

A comparison of modelling options to assess annual variation in susceptibility of generic Escherichia coli isolates to ceftiofur, ampicillin and nalidixic acid from retail chicken meat in Canada

Statistical modelling of antimicrobial resistance (AMR) data is an important aspect of AMR surveillance programs; however, minimum inhibitory concentration (MIC) data can be challenging to model. The conventional approach is to dichotomize data using established breakpoints, then use logistic regression modelling for analysis. A disadvantage of this approach is a loss of information created by dichotomizing the data. The objectives of the study were to compare the performance and results of different regression models for the analysis of annual variation in susceptibility of generic Escherichia coli (E. coli) isolates to ceftiofur, ampicillin and nalidixic acid from retail chicken meat surveillance samples. E. coli susceptibility data for the three antimicrobials from retail chicken samples from 2007 to 2014 were obtained from the Canadian Integrated Program for Antimicrobial Resistance Surveillance (CIPARS). Annual variation in susceptibility for each antimicrobial was evaluated using multivariable linear, tobit, logistic, multinomial, ordinal and complementary log-log regression models (clog-log). MIC (log2), censored MIC (log2), resistant/susceptible, and categorized MIC (3 or 4 categories) data were used as outcome variables for the appropriate statistical models. Year and region were modelled as categorical predictor variables. Random intercepts were included in the ceftiofur and ampicillin models to account for clustering by retail establishment. The model assumptions evaluated for the mixed models included homoscedasticity and normality of residuals (linear and tobit), homoscedasticity and normality of best linear unbiased predictions (all models), proportional odds (ordinal), and proportional hazards (clog-log). Fixed effects models were used for the nalidixic acid models. The model assumptions evaluated for the fixed effects models included homoscedasticity and normality of residuals (linear and tobit), goodness-of-fit test (logistic and multinomial), proportional odds (ordinal), and proportional hazards (clog-log).Only logistic and multinomial models met model assumptions. Significant annual variation in susceptibility to all three antimicrobials was identified by the multinomial regression models, whereas the logistic regression models only identified significant annual variation in susceptibility to ceftiofur. The multinomial regression model consistently identified additional significant annual variation in susceptibility compared to the logistic regression model. The multinomial modelling approach was able to identify differences between MIC categories within susceptible MIC values, which were below the breakpoint (R) detection level. Given the convention of dichotomizing susceptibility data, the logistic regression approach is likely to remain the standard method of analysis for AMR surveillance data; however, the results of this study demonstrate that multinomial regression should be considered for the analysis of AMR surveillance data.