Government Administrative Data Research Articles

HHI: How housing intervention programmes are improving health outcomes for children in New Zealand

ObjectivesIt is estimated that around 30,000 New Zealand (NZ) children are hospitalised each year for preventable conditions related to poor-quality housing. To combat the high levels of housing-related illness in NZ, the Healthy Homes Initiative (HHI) was established. The HHI is a government-funded, multicomponent housing intervention programme operating with the aim of reducing the incidence of housing-related illness in children. The objective of this study was to evaluate the impact of the HHI programme on health outcomes of participating families. ApproachThe Integrated Data Infrastructure (IDI) is a database of linked microdata maintained by Statistics NZ. Deidentified HHI referral data were linked to the IDI, and we used government administrative data to explore the impact of the HHI on health outcomes. ResultsThe HHI dataset linked to the IDI provided referral information for over 21,000 participants of the programme. Characteristics of the cohort included a higher proportion of participants who identified as Māori or Pacific Islander and lower socio-economic status than the general NZ population. Odds ratios were calculated for all variables in the model and showed lower odds of hospitalisation associated with participation in the HHI programme. ConclusionThis study demonstrates how housing-intervention programmes can improve health outcomes for those at risk of illness related to housing. These programmes are imperative for addressing the burden of disease and significant health inequities driven by exposure to substandard housing in NZ. This study shows how data linkage methods and administrative data can be used to evaluate large-scale government housing programmes.

Understanding the developmental well-being of children from refugee backgrounds in British Columbia, Canada: A population-level mixed methods approach

Objectives and ApproachThis study utilized an explanatory sequential mixed methods approach to investigate the developmental well-being of children from refugee backgrounds in British Columbia (BC), Canada. Objective 1 (quantitative) leveraged population-level, government administrative data individually linked to a province-wide, routinely collected, teacher-reported measure of children’s development in kindergarten (the Early Development Instrument; EDI) to examine developmental outcomes across five domains for children identified as first-generation refugees (N=770), first-generation immigrants (N=7875), and non-migrants (N=199,186). In Objective 2 (qualitative), the population-level EDI results were brought to focus groups with BC educators and settlement workers (N=7) who work closely with children from refugee backgrounds to further corroborate, expand, and elaborate on the findings. ResultsA series of multiple linear regression models; adjusted for age, sex, and English Language Learner status showed that first-generation refugee status was significantly predictive of lower EDI scores in the areas of language & cognitive development, communication & generation knowledge, social competence, emotional maturity, and physical health & well-being. Focus group results corroborated the quantitative findings, added critical complexity/context (e.g., impacts of trauma), and identified important policy-oriented levers (e.g., early, accessible assessments and supports). ConclusionsThe study provided an understanding of the population-level developmental well-being of children from refugee backgrounds in BC, framed by rich, contextualized, and actionable knowledge from focus groups. ImplicationsShowcasing the combined breadth and depth of using a mixed methods approach, how we can best support the developmental challenges and build upon the strengths of children from refugee backgrounds will be discussed.

Recommending Career Transitions to Job Seekers Using Earnings Estimates, Skills Similarity, and Occupational Demand

This paper describes a career recommendation algorithm that uses government administrative data to help job seekers discover new careers that similar job seekers have successfully switched to in the past. Algorithm development was motivated by workers and contractors who were displaced by the COVID-19 economic crisis and by workers in declining industries seeking new careers in growing ones. Traditional job boards available through state government websites list all available jobs but do little to remove uncertainty associated with moving to a new industry or occupation. Our recommendation algorithm can lower this uncertainty. It uses causal machine learning techniques and administrative data on the universe of individual-level employment histories and earnings to identify career transitions that have resulted in increased earnings and employment for previous job seekers. We combine these estimates with measures of skill similarity across occupations, derived from natural-language processing of millions of full-text job descriptions, and with occupational demand, as measured by nightly job posting volume. The algorithm parses applicant resumes and returns recommended careers that use similar skills, have available jobs, and are estimated to lead to higher earnings and employment. We have implemented our algorithm in production workforce development systems in five U.S. states.

Do participants in widening participation outreach programmes in England progress to selective universities at a higher rate than would otherwise be expected?

AbstractYoung people growing up in England from a poorer background are less likely to progress into higher education compared to their better off counterparts. This is especially true with respect to more selective universities. This study used government administrative data to gauge the effectiveness of the ‘Realising Opportunities’ programme, which provides a package of support to prospective university applicants from disadvantaged backgrounds to facilitate their progression to more selective universities. Data was gathered concerning 769 16 to 18‐year‐old students who took part in the programme—which is delivered by a consortium of selective universities—between 2015 and 2017. This data revealed participants' personal characteristics (e.g., gender, ethnicity and socioeconomic status), school attainment and eventual university destination (or lack thereof). Additionally, national administrative data was sourced from the Department for Education's National Pupil Database and the Higher Education Statistics Agency for the entire cohort of state school pupils in England who turned 16 in the 2014–2015 academic year. This national data was used to generate a statistical model that could predict the likelihood of a young person in England progressing to a selective university based on their school attainment and other personal characteristics. Data concerning the programme participants was fed through the model and it was observed that the number of participants who had in fact progressed to selective universities greatly exceeded that predicted. Participation in the Realising Opportunities programme therefore appeared to be positively associated with an increased likelihood of progression to a selective university.

Community-led diversion of Indigenous young people from the justice system: The role of government administrative data

Indigenous communities in Australia have grave concerns about the high rates of incarceration of their young people, and advocate for the need for holistic community-led solutions. This article details the use of administrative data in the evaluation of a model of youth diversion led by an Elders group in a remote community, developed in partnership with a university. The model views effective diversion as addressing all underlying factors driving contact with the justice system for Indigenous young people. The methodology includes linking data from relevant government agencies for holistic understanding of young people's institutional engagement over time, including unmet health, housing, education and disability needs and opportunities for positive pathways. Administrative data has not conventionally been subject to ethical scrutiny nor seen as compatible with community priorities. However, we demonstrate how it can enable communities to investigate questions and build change meaningful to them, and improve evidence and policy.

Early Death Among COVID-19 Patients: A Cross-sectional Analysis of the First 10,000 COVID-19 Deaths from the Indian State of Tamil Nadu

BackgroundTamil Nadu state reported the second highest number of confirmed COVID-19 cases in India. In this study, we aimed to describe and determine the risk factors for early death among the first 10,000 COVID-19 deaths in the state.MethodsWe conducted a cross sectional analysis of state government administrative data to describe deaths, examine the differences between early deaths and non-early deaths, and calculate the risks of early death for several independent variables. All p-values < 0.05 were considered statistically significant.ResultsIn total, 4147 early deaths (41.5%) were recorded; the median age of patients who suffered from early death was significantly lower [64 years; interquartile range (IQR): 55–72] when compared with patients who did not suffer from early death (65 years; IQR: 56–73). After adjusting for comorbidities, age, and the time elapsed from the onset of symptoms to hospitalization; we found that the risk of early death was significantly lower for males [adjusted odds ratio (aOR): 0.82; 95% confidence interval (CI): 0.72, 0.93; p = 0.002], among rich individuals (aOR: 0.76; 95% CI: 0.63, 0.92; p = 0.004), in the richest districts (aOR: 0.70; 95% CI: 0.59, 0.84; p < 0.001) and for those who received treatment in private facilities (aOR: 0.45; 95% CI: 0.40, 0.51; p < 0.001.ConclusionsThe risk of early deaths among the first 10,000 reported COVID deaths in the Tamil Nadu state of India was higher in patients treated in government hospitals especially in the poorest districts probably indicating a lack of infrastructure in government facilities or the overburdening of government facilities at least in the early phase of the pandemic.

Semantics-Aware Document Retrieval for Government Administrative Data

The process of data analytics on large-scale government administrative data — that belong to various domains like education, transport, energy, and health — can be enhanced by retrieving pertinent documents from diverse data sources. Without a supporting framework of metadata, big data analytics can be daunting. Even though statistical algorithms can perform extensive analyses on a variety of data with little help from metadata, applying these techniques to heterogeneous data may not always result in reliable findings. Recently, semantics-aware (or semantic search) search techniques received much attention as they utilize implicit knowledge to enhance the search. Similarly, traditional search engines rely on the inherent linkages within the underlying data model to improve their search quality. In the case of general-purpose information retrieval systems, to gather information from the internet (open access data) or to access open government administrative data, a domain agnostic ontology shall be employed to supply background knowledge. This paper draws on research undertaken by the authors at IIIT Bangalore Center for Open Data Research (CODR) in developing a semantics-aware data lake framework to host and analyze government administrative data. In this study, we present an ontology-based document retrieval solution where an ontology serves as an intermediary to close the gap between what the user seeks and what the search retrieves. Although our study settings are based on the Government of Karnataka (GoK, India), we believe the findings have wider resonance. Our experimental results based on agricultural data from the GoK look promising.

Opportunities and Challenges for Official Statistics in Digital Bangladesh

Governments, businesses, and the public, in general, are expected to receive information from national systems of official statistics regarding the economic, demographic, social, and environmental situation. Therefore, digitizing data collection for government statistics could have a significant social impact. Slowly but surely, government administrative data is being used more. There is scant evidence, but significant data access through satellites, point-of-sale systems, and social networks is now being researched and tested. New needs for official statistics and pressures on traditional data collection from households and companies drove these projects. New requirements for official statistics have brought about these requirements and demands. The efficient use of new data sources by the government can lay the most important evidential foundation for policies that are intended to enhance people's lives, and data that has been anonymized is also being made available to other researchers. The study aims to explore the opportunities and challenges for official statistics in digital Bangladesh and identify the potential benefits and drawbacks of using digital technologies and data sources for official statistics. By understanding these opportunities and challenges, policymakers, statisticians, and other stakeholders can develop strategies and solutions that address the challenges and leverage the opportunities to improve official statistics' quality, timeliness, and relevance. Additionally, the objective is to highlight the importance of investing in capacity building, data governance frameworks, and technology and infrastructure to promote the development of official statistics in digital Bangladesh.

Association between children's health-related quality of life, healthcare costs and socioeconomic position: Results from a longitudinal Australian-based study

BackgroundEquity in healthcare service access and use are national goals and principal indicators of health system performance. Whilst it is important to understand how healthcare costs, socioeconomic position (SEP), and children's well-being are associated, limited literature explores how these covary. AimsTo explore the associations between families' SEP and (1) children's health-related quality of life (HRQoL) from 4 to 13 years; and (2) healthcare costs from 0 to 13 years; as well as (3) contributing factors that influence children's HRQoL and healthcare costs. MethodsData were drawn from the Longitudinal Study of Australian Children (LSAC). We used the parent-reported Pediatric Quality of Life Inventory 4.0 to measure children's HRQoL from 4 to 13 years. Healthcare costs to government from 0 to 13 years were measured using the government administrative data (Medicare) which includes records of both medical and pharmaceutical service utilisation. We used a composite of income, parental education, and occupation to measure SEP. We used linear and mixed effects modelling to explore associations between families' SEP and children's HRQoL or government healthcare costs, including the effects of predictors for these associations. Analyses were weighted and accounted for the survey design. ResultsCompared to children from high SEP families, those from low SEP families had lower HRQoL (from age 6–13 years, p = 0.001) and their families incurred higher government healthcare costs (from birth to 13 years). This association was confirmed using the mixed effects model (p = 0.054). Child special healthcare needs, being a single-parent or having parental stress were related to poorer children's HRQoL and higher government healthcare costs. Living in regional and remote areas was related to lower government healthcare costs. ConclusionUp to 13 years of age, children with low SEP used more health services but had lower HRQoL than those from high SEP families. These findings highlight a need to support children from low SEP families to improve their health and wellbeing.

Trade liberalization and firm toxic emissions

AbstractWe investigate the effects of trade liberalization on firm‐level toxic emissions. Using China's World Trade Organization (WTO) accession as an exogenous shock, we further introduce government administrative data of firm toxic emissions to conduct a difference‐in‐differences estimation. We show that trade liberalization significantly reduces firm toxic emissions. Our results are robust to alternative measures and different specifications. Endogeneity tests based on different empirical designs or omitted variables further validate our main findings. China's WTO accession primarily affected pollution emissions through the scale of production, export opportunities, technology upgrading, intermediate inputs, and within‐firm responses mechanisms.

A Novel Imputation Approach for Sharing Protected Public Health Data.

Objectives. To develop an imputation method to produce estimates for suppressed values within a shared government administrative data set to facilitate accurate data sharing and statistical and spatial analyses. Methods. We developed an imputation approach that incorporated known features of suppressed Massachusetts surveillance data from 2011 to 2017 to predict missing values more precisely. Our methods for 35 de-identified opioid prescription data sets combined modified previous or next substitution followed by mean imputation and a count adjustment to estimate suppressed values before sharing. We modeled 4 methods and compared the results to baseline mean imputation. Results. We assessed performance by comparing root mean squared error (RMSE), mean absolute error (MAE), and proportional variance between imputed and suppressed values. Our method outperformed mean imputation; we retained 46% of the suppressed value's proportional variance with better precision (22% lower RMSE and 26% lower MAE) than simple mean imputation. Conclusions. Our easy-to-implement imputation technique largely overcomes the adverse effects of low count value suppression with superior results to simple mean imputation. This novel method is generalizable to researchers sharing protected public health surveillance data. (Am J Public Health. 2021; 111(10):1830-1838. https://doi.org/10.2105/AJPH.2021.306432).

844Considering child maltreatment in social inequalities of educational achievement: a whole-of-population data linkage study

Abstract Background Strategies aimed at reducing health inequalities include a focus on improving education, a key determinant of health and labour market outcomes. Studies have demonstrated that both early life socioeconomic disadvantage and child maltreatment result in poorer educational outcomes, but have not examined the effect of the combination of these adversities. Methods This study used de-identified, linked government administrative data from the South Australian Early Childhood Data Project on children born in South Australia who completed Year 5 National Assessment Program—Literacy and Numeracy (NAPLAN) during 2008-2015 (n = 61,445), and their corresponding birth, perinatal, school enrolment and child protection system data. Results Overall, 23.7% (95% CI 23.4-24.1) had a poor NAPLAN outcome (at or below national minimum standard on three or more of the five domains - reading, writing, spelling, grammar and punctuation, numeracy). This proportion was approximately double for children who experienced substantiated maltreatment compared to children who did not experience contact with the child protection system, irrespective of the social circumstances (parental education, employment, housing and area-level disadvantage) in which the child lived. Conclusions Social advantage is not sufficient to protect children from poor educational outcomes if they experience substantiated maltreatment. Interventions and services will need to focus on child maltreatment, in addition to socioeconomic factors, to improve educational outcomes in the population. Key messages Within each social group, from advantaged to disadvantaged children, the risk of poor NAPLAN outcomes was greater for children who had contact with the child protection system.

838Thriving in adversity: positive child development despite early disadvantage in a whole-of-population data linkage study

Abstract Background Experiencing socioeconomic disadvantage, poor health, or child maltreatment in early life has negative effects on child development. However, we know little about children who have good developmental outcomes despite experiencing adversity. Methods This study used de-identified, linked government administrative data from the South Australian Early Childhood Data Project: specifically Australian Early Development Census (AEDC) data for all South Australian born children in their first year of school in 2009, 2012 and 2015 (n = 47,179) and their corresponding birth, perinatal, school enrolment, hospital admission, emergency department presentation, public housing and child protection data. Latent class analyses constructed multidimensional measures of socioeconomic, health, and maltreatment adversities experienced from birth to age 5. Results Overall, 49.8% (95% CI 49.2-50.4) of children were on track on all five AEDC domains, but this ranged from 53.7% among children who did not experience high levels of adversity to 13.5% among children with high levels of all three adversities. Conclusions Among children who experienced high levels of two or three early adversity types, approximately 1 in 5 were developmentally on track. Understanding characteristics of these children who thrive, against the odds, will help identify intervention opportunities to improve child development. Key messages Compared with children who did not experience high levels of adversity, each additional adversity reduced the likelihood of being developmentally on track by approximately 10% to 15%. Children experiencing socioeconomic or maltreatment adversity were less likely to be developmentally on track than children experiencing health adversity.

Sharing administrative health data with private industry: A report on two citizens' juries.

BackgroundThere is good evidence of both community support for sharing public sector administrative health data in the public interest and concern about data security, misuse and loss of control over health information, particularly if private sector organizations are the data recipients. To date, there is little research describing the perspectives of informed community members on private sector use of public health data and, particularly, on the conditions under which that use might be justified.MethodsTwo citizens' juries were held in February 2020 in two locations close to Sydney, Australia. Jurors considered the charge: ‘Under what circumstances is it permissible for governments to share health data with private industry for research and development?’ResultsAll jurors, bar one, in principle supported sharing government administrative health data with private industry for research and development. The support was conditional and the juries' recommendations specifying these conditions related closely to the concerns they identified in deliberation.ConclusionThe outcomes of the deliberative processes suggest that informed Australian citizens are willing to accept sharing their administrative health data, including with private industry, providing the intended purpose is clearly of public benefit, sharing occurs responsibly in a framework of accountability, and the data are securely held.Patient and Public ContributionThe design of the jury was guided by an Advisory Group including representatives from a health consumer organization. The jurors themselves were selected to be descriptively representative of their communities and with independent facilitation wrote the recommendations.

The feasibility and challenge of using administrative data: a case study of historical prisoner surveys

ABSTRACT The benefits of government administrative data are well-established, but research has not examined processes by which these data can be identified, organised and governed in line with ethical considerations. This is an important undertaking given the potential benefits of robust administrative data, especially at a time when the collection of primary data has been made more challenging due to COVID-19 restrictions. We use principles of Big Data to characterise a large operational data set comprised of responses from a prisoner survey about experiences of detainment in England and Wales. These findings are discussed in relation to future research with administrative data and the value that historical survey data can provide to various beneficiaries.

Using web content analysis to create innovation indicators—What do we really measure?

This study explores the use of web content analysis to build innovation indicators from the complete texts of 79 corporate websites of Canadian nanotechnology and advanced materials firms. Indicators of four core concepts (R&amp;D, IP protection, collaboration, and external financing) of the innovation process were built using keywords frequency analysis. These web-based indicators were validated using several indicators built from a classic questionnaire-based survey with the following methods: correlation analysis, multitraits multimethods (MTMM) matrices, and confirmatory factor analysis (CFA). The results suggest that formative indices built with the questionnaire and web-based indicators measure the same concept, which is not the case when considering the items from the questionnaire separately. Web-based indicators can act either as complements to direct measures or as substitutes for broader measures, notably the importance of R&amp;D and the importance of IP protection, which are normally measured using conventional methods, such as government administrative data or questionnaire-based surveys.

Comparing Reports of Child Sexual and Physical Abuse Using Child Welfare Agency Data in Two Jurisdictions with Different Mandatory Reporting Laws

Empirical analysis has found that mandatory reporting legislation has positive effects on case identification of child sexual abuse both initially and over the long term. However, there is little analysis of the initial and ongoing impact on child protection systems of the rate of reports that are made if a reporting duty for child sexual abuse is introduced, especially when compared with rates of reports for other kinds of child maltreatment. This research analysed government administrative data at the unique child level over a seven-year period to examine trends in reports of child sexual abuse, compared with child physical abuse, in two Australian states having different socio-legal dimensions. Data mining generated descriptive statistics and rates per 100,000 children involved in reports per annum, and time trend sequences in the seven-year period. The first state, Western Australia, introduced the legislative reporting duty in the middle of the seven-year period, and only for sexual abuse. The second state, Victoria, had possessed mandatory reporting duties for both sexual and physical abuse for over a decade. Our analysis identified substantial intra-state increases in the reporting of child sexual abuse attributable to the introduction of a new legislative reporting duty, and heightened public awareness resulting from major social events. Victoria experienced nearly three times as many reports of physical abuse as Western Australia. The relative burden on the child protection system was most clearly different in Victoria, where reports of physical abuse were relatively stable and two and a half times higher than for sexual abuse. Rates of children in reports, even at their single year peak, indicate sustainable levels of reporting for child welfare agencies. Substantial proportions of reports were made by both legislatively mandated reporters, and non-mandated community members, suggesting that government agencies would benefit from engaging with communities and professions to enhance a desirable reporting practice.

Area-Level Variation in Children's Unmet Need for Community-Based Mental Health Services: Findings from the 2014 Ontario Child Health Study.

There is limited empirical evidence documenting the magnitude and correlates of area-level variability in unmet need for children's mental health services. Research is needed that identifies area-level characteristics that can inform strategies for reducing unmet need in the population. The study purpose is to: (1) estimate area-level variation in children's unmet need for mental health services (using Service Areas as defined by the Ontario Ministry of Children and Youth Services), and (2) identify area-level service arrangements, and geographic and population characteristics associated with unmet need. Using individual-level general population data, area-level government administrative data and Census data from Ontario, Canada, we use multilevel regression models to analyze unmet need for mental health services among children (level 1) nested within Service Areas (level 2). The study finds that 1.64% of the reliable variance in unmet need for mental health services is attributable to between-area differences. Across areas, we find that Service Areas with more agencies had a lower likelihood of unmet need for mental health services. Compared to other Service Areas, Toronto had much lower likelihood of unmet need compared to the rest of Ontario. Rural areas, areas with unsatisfactory public transport, and areas with higher levels of socio-economic disadvantage had a higher likelihood of unmet need for mental health services. These findings identify challenges in service provision that researchers, policymakers and administrators in children's mental health services need to better understand. Policy implications and potential Service Area strategies that could address equitable access to mental health services are discussed.

Predicting high-risk opioid prescriptions before they are given

Misuse of prescription opioids is a leading cause of premature death in the United States. We use state government administrative data and machine learning methods to examine whether the risk of future opioid dependence, abuse, or poisoning can be predicted in advance of an initial opioid prescription. Our models accurately predict these outcomes and identify particular prior nonopioid prescriptions, medical history, incarceration, and demographics as strong predictors. Using our estimates, we simulate a hypothetical policy which restricts new opioid prescriptions to only those with low predicted risk. The policy's potential benefits likely outweigh costs across demographic subgroups, even for lenient definitions of "high risk." Our findings suggest new avenues for prevention using state administrative data, which could aid providers in making better, data-informed decisions when weighing the medical benefits of opioid therapy against the risks.

Harnessing administrative data for humanitarian responses

The need to improve data to support effective responses in humanitarian settings and emergencies is widely recognized; and there is a growing understanding that administrative data systems can be an important source of data in these contexts. Government data systems however are often overlooked, particularly in developing countries, due to access or quality concerns. New systems are frequently established to respond to humanitarian data needs – particularly in protracted emergency settings. This may be necessary in some circumstances due to a collapse or absence of existing systems, but often duplicates government information structures. In the long term, parallel systems risk undermining national systems which may take a significant time to recover or re-build, and can divert highly limited resources.&#x0D; Routine government administrative data systems need to be both more responsive to changes in the environments and communities in which they are operating, and resilient enough to remain operational and secure in those changed circumstances. This work aims to explore what this means in system design, operation, governance, and preparedness for key sectors; to provide practical guidance for national governments and supporting agencies.&#x0D; Findings will be drawn from literature review, expert opinion, and field examples. Case studies will be undertaken in two countries that are either considered “fragile” or have recently responded to an emergency. Work will focus on areas in which UNICEF has a formal role, however it is anticipated to highlight issues relevant to a broad spectrum of humanitarian data needs.&#x0D; Preliminary findings are anticipated to be published towards the end of 2019; and will compliment work programs on developing an understanding of administrative data maturity, and the global standards and norms for upholding the rights of children affected by humanitarian crises, such as UNICEF's Core Commitments for Children in Humanitarian Action.