A semi-Markov multistate cure model for estimating intervention effects in stepped wedge design trials.

Multistate models are useful for studying exposures that affect transitions among a set of health states. However, they can be challenging to apply when exposures are time-varying. We develop a multistate model and a method of likelihood construction that allows application of the model to data in which interventions or other exposures can be time-varying and an individual may to be exposed to multiple intervention conditions while progressing through states. The model includes cure proportions, reflecting the possibility that some individuals will never leave certain states. We apply the approach to analyze patient vaccination data from a stepped wedge design trial evaluating two interventions to increase uptake of human papillomavirus vaccination. The states are defined as the number of vaccine doses the patient has received. We model state transitions as a semi-Markov process and include cure proportions to account for individuals who will never leave a given state (e.g. never receive their next dose). Multistate models typically quantify intervention effects as hazard ratios contrasting the intensities of transitions between states in intervention versus control conditions. For multistate processes, another clinically meaningful outcome is the change in the percentage of the study population that has achieved a specific state (e.g. completion of all required doses) by a specific point in time due to an intervention. We present a method for quantifying intervention effects in this manner. We apply the model to both simulated and real-world data and also explore some conditions under which such models may give biased results.

Impact of a Multi-Level, Multi-Component, System Intervention on HPV Vaccination in a Federally Qualified Health Center.

BACKGROUND: Human papillomavirus (HPV) vaccines can significantly reduce the burden of HPV-associated cancers, but remain underutilized. We evaluated a multi-component, system-level intervention to improve HPV vaccination in a large Federally Qualified Health Center (FQHC) that serves a primarily low income Latino population. METHODS: From January 2015 through March 2017, we evaluated the effectiveness of a multi-component, system-level intervention to improve HPV vaccination rates in eight clinics randomly assigned to study condition (four intervention, four usual care). The intervention included parent reminders for HPV vaccine series completion, provider training, clinic-level audit and feedback, and workflow modifications to reduce missed opportunities for vaccination. Using a difference-in-differences approach, we compared HPV vaccination rates among patients, ages 11 to 17 during a 12-month preintervention period and a 15-month intervention period. Linear mixed models were used to estimate intervention effects on vaccine initiation and completion. RESULTS: The sample included approximately 15,000 adolescents each quarter (range 14,773-15,571; mean age 14 years; 51% female, 88% Latino). A significantly greater quarterly increase in HPV vaccine initiation was observed for intervention compared with usual care clinics (0.75 percentage point greater increase, P < 0.001), corresponding to 114 additional adolescents vaccinated per quarter. The intervention led to a greater increase in HPV vaccine completion rates among boys (0.65 percentage point greater increase, P < 0.001), but not girls. CONCLUSIONS: Our system-level intervention was associated with modest improvements in HPV vaccine initiation overall and completion among boys. IMPACT: Study findings have implications for reducing HPV-related cancers in safety net populations.

Power analysis for stepped wedge trials with multiple interventions.

Stepped wedge design (SWD) trials are cluster randomized trials that feature staggered, unidirectional cross-over between treatment conditions. Existing literature on power for SWDs focuses primarily on designs with two conditions, typically a control and an intervention condition. However, SWDs with more than one treatment condition are being proposed and conducted. We present a linear mixed model for SWDs with two or more interventions, including both multiarm and factorial designs. We derive standard errors of the intervention effect coefficients, and present power calculation methods. We consider both repeated cross-sectional and cohort designs. Design features, with a focus on treatment allocations, are examined to determine their impact on power.

Family disability, poverty and parenting stress: Analysis of a cross-sectional study in Kenya.

BACKGROUND: Households with a disabled member, be they a caregiver or a child, are poorer than households not affected by disability. Poverty, caregiving as a person with a disability and being the caregiver of a child with a disability can lead to increased parenting stress. OBJECTIVES: The objective of this study was to examine whether parenting stress experienced by caregivers in a household with a disabled member is greater when the disabled member is the caregiver, or the child, and how much of these respective relationships is explained by poverty. METHOD: We collected cross-sectional data using a demographic survey, the Washington Group Questions on adult disability, the 10 Questions on child disability and the Parenting Stress Index-Short Form, from 465 caregivers enrolled in a non-governmental child development programme in Kenya. RESULTS: Households with a disabled member were poorer than households without a disabled member. Parenting stress of disabled caregivers was higher than parenting stress of non-disabled caregivers; however, this relationship disappeared when socio-economic status was controlled for. Caregivers of disabled children were more stressed than caregivers of non-disabled children, and this effect was not explained by differences in socio-economic status. CONCLUSION: Our findings highlight the importance of developing a comprehensive understanding of the stressors facing households with a disabled member, particularly if that member is a child, so that supportive interventions can adequately cater to the needs of caregivers, and their children, in the context of poverty.

Pandemic velocity: Forecasting COVID-19 in the US with a machine learning & Bayesian time series compartmental model.

Predictions of COVID-19 case growth and mortality are critical to the decisions of political leaders, businesses, and individuals grappling with the pandemic. This predictive task is challenging due to the novelty of the virus, limited data, and dynamic political and societal responses. We embed a Bayesian time series model and a random forest algorithm within an epidemiological compartmental model for empirically grounded COVID-19 predictions. The Bayesian case model fits a location-specific curve to the velocity (first derivative) of the log transformed cumulative case count, borrowing strength across geographic locations and incorporating prior information to obtain a posterior distribution for case trajectories. The compartmental model uses this distribution and predicts deaths using a random forest algorithm trained on COVID-19 data and population-level characteristics, yielding daily projections and interval estimates for cases and deaths in U.S. states. We evaluated the model by training it on progressively longer periods of the pandemic and computing its predictive accuracy over 21-day forecasts. The substantial variation in predicted trajectories and associated uncertainty between states is illustrated by comparing three unique locations: New York, Colorado, and West Virginia. The sophistication and accuracy of this COVID-19 model offer reliable predictions and uncertainty estimates for the current trajectory of the pandemic in the U.S. and provide a platform for future predictions as shifting political and societal responses alter its course.

Associations between caregiver mental health and young children's behaviour in a rural Kenyan sample.

Background: Research shows that caregiver mental health problems have direct, significant effects on child behaviour. While these risks are amplified in low-resource settings, limited evidence exists from these places, especially sub-Saharan Africa. Objective: We measured associations between caregiver mental health and child behaviour in a rural Kenyan sample, hypothesizing that higher rates of caregiver mental health would be associated with increased child behavioural problems. We also sought to provide an overview of caregiver mental health symptoms in our sample. Method: Cross-sectional data were collected from caregivers of children ages 4-5 years old enrolled in a community-based early child development programme in western Kenya. 465 caregivers were recruited and assessed at baseline, and answered questions about child behaviour, mental health symptoms (depression, anxiety, stress), and help-seeking. A multivariate linear regression model was used to assess significance of each mental health factor. Results: Caregiver anxiety (p = 0.01) and parenting stress (p < 0.001) were significantly associated with child behavioural problems. 245 caregivers (52.9%) had high levels of symptoms of depression, anxiety, or both; furthermore, 101 caregivers (21.7%) scored above the cut-off for both of these scales. A high proportion of our sample (60.6%) reported seeking some formal or informal psychosocial support services; however, less than one-third of these caregivers were symptomatic (30.9%). Conclusion: Anxiety and stress were associated with poorer child behavioural outcomes. Our sample reflected a higher prevalence of caregiving adults with mental health symptomology than previous estimates from Kenya, with few high-symptom caregivers seeking support. We discuss further implications for programming and health services delivery.

Pseudo-likelihood based logistic regression for estimating COVID-19 infection and case fatality rates by gender, race, and age in California.

In emerging epidemics, early estimates of key epidemiological characteristics of the disease are critical for guiding public policy. In particular, identifying high-risk population subgroups aids policymakers and health officials in combating the epidemic. This has been challenging during the coronavirus disease 2019 (COVID-19) pandemic because governmental agencies typically release aggregate COVID-19 data as summary statistics of patient demographics. These data may identify disparities in COVID-19 outcomes between broad population subgroups, but do not provide comparisons between more granular population subgroups defined by combinations of multiple demographics. We introduce a method that helps to overcome the limitations of aggregated summary statistics and yields estimates of COVID-19 infection and case fatality rates - key quantities for guiding public policy related to the control and prevention of COVID-19 - for population subgroups across combinations of demographic characteristics. Our approach uses pseudo-likelihood based logistic regression to combine aggregate COVID-19 case and fatality data with population-level demographic survey data to estimate infection and case fatality rates for population subgroups across combinations of demographic characteristics. We illustrate our method on California COVID-19 data to estimate test-based infection and case fatality rates for population subgroups defined by gender, age, and race/ethnicity. Our analysis indicates that in California, males have higher test-based infection rates and test-based case fatality rates across age and race/ethnicity groups, with the gender gap widening with increasing age. Although elderly infected with COVID-19 are at an elevated risk of mortality, the test-based infection rates do not increase monotonically with age. The workforce population, especially, has a higher test-based infection rate than children, adolescents, and other elderly people in their 60-80. LatinX and African Americans have higher test-based infection rates than other race/ethnicity groups. The subgroups with the highest 5 test-based case fatality rates are all-male groups with race as African American, Asian, Multi-race, LatinX, and White, followed by African American females, indicating that African Americans are an especially vulnerable California subpopulation.

Disparities in cancer mortality in Los Angeles County, 1999-2013: an analysis comparing trends in under-resourced and affluent regions.

PURPOSE: While cancer mortality has declined by 27% between 1991 and 2016 in the United States, there are large disparities in cancer mortality by racial/ethnic groups, socioeconomic status and access to care. The purpose of this analysis is to compare trends in cancer mortality among regions (Service Planning Areas, SPAs) in Los Angeles (LA) County that vary with respect to racial/ethnic distribution and social determinants of health, including poverty, education and access to care. METHODS: We estimated age- and race/ethnicity-standardized mortality for lung, colorectal (CRC) and breast cancer for eight SPAs from 1999 to 2013. We calculated three recommended measures of disparities that reflect absolute, relative and between-group disparities. RESULTS: In all of LA County, statistically significant declines in age- and race/ethnicity-standardized mortality ranged from 30% for lung cancer to 20% for CRC to 15% for breast cancer. Despite some of the largest declines in the most under-resourced SPAs (South LA, East LA, South Bay), disparities between the lowest and highest mortality by SPA did not significantly change from 1999 to 2013. CONCLUSIONS: Despite significant declines in cancer mortality in LA County from 1999 to 2013, and in racial/ethnic groups, there was little progress toward reducing disparities among SPAs. Highest mortalities for the three cancers were observed in Antelope Valley, San Fernando Valley, San Gabriel Valley, South LA and East LA. Findings demonstrate the importance of examining regional differences in cancer mortality to identify areas with highest needs for interventions and policies to reduce cancer disparities.

Associations between young children's exposure to household violence and behavioural problems: Evidence from a rural Kenyan sample.

Little is known about how young children in low- and middle-income countries (LMICs) experience violence in their homes, and how different types of household violence may affect child development. This study reports on levels of exposure to household violence and associations with child behavioural outcomes in preschool-aged children in western Kenya. A sample of 465 caregivers, whose children (n = 497) attended early learning centres supported by an international NGO, were enrolled in the study. Caregivers reported on exposure to intimate partner violence (IPV), household discipline practices, attitudes about gender roles, and child behavioural outcomes. Multivariable analysis showed significant predictive effects of IPV (regression coefficient = 1.35, SE = 0.54, p = 0.01) and harsh psychological child discipline (regression coefficient = 0.74, SE = 0.22, p = 0.001), but not physical discipline (regression coefficient = 0.42, SE = 0.24, p = 0.08), on worse child behavioural problems. These findings indicate that child exposure to violence in different forms is highly prevalent, and associated with poorer outcomes in young children. Community-based programmes focused on parenting and early child development are well-positioned to address household violence in LMIC settings, but must be supported to provide a broader understanding of violence and its immediate and long-term consequences.

Why do entrepreneurial mHealth ventures in the developing world fail to scale?

Telemedicine is an increasingly common approach to improve healthcare access in developing countries with fledgling healthcare systems. Despite the strong financial, logistical and clinical support from non-governmental organisations (NGOs), government ministries and private actors alike, the majority of telemedicine projects do not survive beyond the initial pilot phase and achieve their full potential. Based on a review of 35 entrepreneurial telemedicine and mHealth ventures, and 17 reports that analyse their operations and challenges, this article provides a narrative review of recurring failure modes, i.e. factors that lead to failure of such venture pilots. Real-world examples of successful and failed ventures are examined for key take-away messages and practical strategies for creating commercial viable telemedicine operations. A better understanding of these failure modes can inform the design of sustainable and scalable telemedicine systems that effectively address the growing healthcare disparities in developing countries.