Household Health Care Payments Under Rate Setting, Spending Growth Target, and Single-Payer Policies.

Importance: Households have high burden of health care payments. Alternative financing approaches could reduce this burden for some households. Objective: To estimate the distribution of household health care payments across income under health care reform policies. Design, Setting, and Participants: Cross-sectional study with microsimulation used nationally representative data of the US population in 2030. Civilian, noninstitutionalized population from the 2022 Current Population Survey linked to expenditures from the 2018 and 2019 Medical Expenditure Panel Survey and 2022 National Health Expenditure Accounts were included. Exposure: Rate regulation of hospital, physician, and other health care professional payments equal to the all-payer mean in the status quo, spending growth target at 4% annual per capita growth, and single-payer health care financed through taxes. Main Outcomes and Measures: Household health care payments (out-of-pocket expenses, premiums, and taxes) as a share of compensation. Results: The synthetic population contained 154 456 records representing 339.5 million individuals, with 51% female, 7% Asian, 14% Black, 18% Hispanic White, 56% non-Hispanic White, and 5% other races and ethnicities (American Indian or Alaskan Native only; Native Hawaiian or other Pacific Islander only; and 2 or more races). In the status quo, mean household health care payments as a share of compensation was 24% to 27% (standard error [SE], 0.2%-1.2%) across income groups (median [IQR] 22% [4%-52%] below 139% of the federal poverty level [FPL]; 21% [4%-34%] for households above 1000% FPL [11% of the population]). Under rate setting, mean (SE) payments by households above 1000% FPL increased to 29% (0.6%) (median [IQR], 22% [6%-35%]) and decreased to 23% to 25% for other income groups. Under the spending growth target, mean (SE) payments decreased from 23% to 26% (SE, 0.2%-1.2%) across income groups. Under the single-payer system, mean (SE) payments declined to 15% (0.7%) (median [IQR], 4% [0%-30%]) for those below 139% FPL and increased to 31% (0.6%) (median [IQR], 23% [3%-39%]) for those above 1000% FPL. Uninsurance fell from 9% to 6% under rate setting due to improved Medicaid access, and to zero under the single-payer system. Conclusions and Relevance: Single-payer financing based on the current federal income tax schedule and a payroll tax could substantially increase progressivity of household payments by income. Rate setting led to slight increases in payments by higher-income households, who financed higher payment rates in Medicare and Medicaid. Spending growth targets reduced payments slightly for all households.

Assessing the Impact of Individual Market Reforms in Minnesota.

Starting in 2026, Minnesota could experience disruptions to its health insurance marketplace caused by the anticipated sunset of federal premium subsidy enhancements, made available through the Inflation Reduction Act of 2022, as well as the expiration of state funding for its reinsurance program. With reduced premium subsidies, fewer people might enroll in marketplace plans, which could lead to higher premiums and market instability. The expiration of reinsurance, which partially offsets insurers' claims costs for people with high expenditures, could exacerbate these issues. In this study, researchers estimate the effects of implementing state-funded subsidies to bolster Minnesota's marketplace given these anticipated changes. They also study the impact of replacing the state's Basic Health Program with a similarly structured marketplace plan. The policy reforms that researchers consider were developed by the Minnesota Council of Health Plans and share similar goals with legislation recently proposed by Minnesota policymakers, such as HF 96, a bill authorizing study of a public option that also proposed to temporarily enhance marketplace subsidies.

Strategies to Improve the Impact of Artificial Intelligence on Health Equity: Scoping Review.

BACKGROUND: Emerging artificial intelligence (AI) applications have the potential to improve health, but they may also perpetuate or exacerbate inequities. OBJECTIVE: This review aims to provide a comprehensive overview of the health equity issues related to the use of AI applications and identify strategies proposed to address them. METHODS: We searched PubMed, Web of Science, the IEEE (Institute of Electrical and Electronics Engineers) Xplore Digital Library, ProQuest U.S. Newsstream, Academic Search Complete, the Food and Drug Administration (FDA) website, and ClinicalTrials.gov to identify academic and gray literature related to AI and health equity that were published between 2014 and 2021 and additional literature related to AI and health equity during the COVID-19 pandemic from 2020 and 2021. Literature was eligible for inclusion in our review if it identified at least one equity issue and a corresponding strategy to address it. To organize and synthesize equity issues, we adopted a 4-step AI application framework: Background Context, Data Characteristics, Model Design, and Deployment. We then created a many-to-many mapping of the links between issues and strategies. RESULTS: In 660 documents, we identified 18 equity issues and 15 strategies to address them. Equity issues related to Data Characteristics and Model Design were the most common. The most common strategies recommended to improve equity were improving the quantity and quality of data, evaluating the disparities introduced by an application, increasing model reporting and transparency, involving the broader community in AI application development, and improving governance. CONCLUSIONS: Stakeholders should review our many-to-many mapping of equity issues and strategies when planning, developing, and implementing AI applications in health care so that they can make appropriate plans to ensure equity for populations affected by their products. AI application developers should consider adopting equity-focused checklists, and regulators such as the FDA should consider requiring them. Given that our review was limited to documents published online, developers may have unpublished knowledge of additional issues and strategies that we were unable to identify.

Expanding Insurance Coverage to Undocumented Immigrants in Connecticut.

Policymakers in Connecticut are considering various state-funded policy options to improve insurance coverage among undocumented and legally present recent immigrants in the state - almost 60 percent of whom lack health insurance. In particular, they are removing immigration status requirements from Medicaid eligibility. They are also considering whether to provide state-funded subsidies to undocumented immigrants enrolled in individual market plans. A key challenge for this analysis was determining what share of undocumented immigrants would be likely to take up insurance coverage if it were available to them. Because few states have expanded coverage to their undocumented populations and because the denominator is uncertain, estimates of take-up rates are highly uncertain. There is similar uncertainty in estimating how much health care undocumented populations will use once they become insured. To address these uncertainties, the authors conducted sensitivity analyses that varied both the take-up and utilization rates. Using the RAND Corporation's COMPARE microsimulation model, the authors estimate the impacts of each policy scenario on enrollment, premiums, state spending, and hospital spending on uncompensated care. Their analysis suggests that removing immigration status requirements for Medicaid and individual market subsidy eligibility would decrease uninsurance among the undocumented and legally present recent immigrant populations by 32 to 37 percent and could improve insurance coverage and affordability in Connecticut for these populations while not substantially impacting other Connecticut residents.

Increasing Subsidies and Expanding Health Insurance Options in Connecticut.

Policymakers in Connecticut are considering various options to increase the affordability of insurance in the state, such as expansions to premium and cost-sharing reduction subsidies on the state's health insurance marketplace, as well as expanded plan offerings, including extending eligibility for the state employee health plan (SEHP) to other groups and a publicly contracted, privately operated plan (the public option plan) offered to individuals on the marketplace. The authors used the RAND Corporation's COMPARE microsimulation model to estimate the impacts of such policy options. For each policy scenario, they calculated enrollment, premiums, consumer spending, and state spending and considered whether the results differed by race, ethnicity, or income group. The individual market reforms substantially increased affordability for people with incomes between 175 and 200 percent of the federal poverty level (FPL), reducing out-of-pocket spending as a share of income by 50 percent in some scenarios. Changes to affordability for higher-income groups were smaller, in part because the proposed policy changes for people with incomes between 200 and 400 percent of FPL were relatively modest and focused only on reducing cost-sharing (not premiums). New costs to the state for 2023 ranged from $19 million to $94 million, depending on the scenario. All four SEHP specifications led to the same bottom-line conclusion that offering a SEHP plan would improve insurance coverage and affordability for those eligible for the plan. Expanding eligibility for the SEHP holds promise for stabilizing or reducing consumer costs, improving plan generosity, and bringing more people into the market.

Summary of the Current Status of Health Insurance Enrollment in Connecticut.

The state of Connecticut is considering a number of policy options to improve health insurance affordability, access, and equity. To create policies designed to increase insurance coverage and access to care in underserved communities and reduce racial and ethnic disparities, state policymakers need an accurate picture of the current distributions of insurance enrollment across these dimensions. The authors combine data from the American Community Survey Public Use Microdata Sample, which includes demographic characteristics, as well as insurance status, with various data sources from the state to provide a fuller picture of insurance enrollment among those under the age of 65 in Connecticut. They also use existing high-level estimates of 2020 insurance enrollment to provide estimates of how enrollment in the state was affected during the early months of the pandemic. The authors find that insurance enrollment in Connecticut in 2019 was generally high but that there were substantial differences in insurance coverage by race and ethnicity. Asian individuals had the highest rates of employer-sponsored insurance coverage, and Black individuals had the highest rates of Medicaid coverage. Hispanic individuals had a higher rate of Medicaid coverage than non-Hispanic individuals. High-level estimates of changes in insurance coverage during the early months of the COVID-19 pandemic suggest that uninsurance decreased slightly, Medicaid coverage increased, and private insurance coverage fell. This study provides the state of Connecticut with estimates of enrollment in detailed health insurance categories by age, gender, race, and ethnicity and highlights the need for better, more-detailed health insurance enrollment data.

INFLATE: a protocol for a randomised controlled trial comparing nasal balloon autoinflation to no nasal balloon autoinflation for otitis media with effusion in Aboriginal and Torres Strait Islander children.

BACKGROUND: Otitis media with effusion (OME) is common and occurs at disproportionately higher rates among Indigenous children. Left untreated, OME can negatively affect language, development, learning, and health and wellbeing throughout the life-course. Currently, OME care includes observation for 3 months followed by consideration of surgical ventilation tube insertion. The use of a non-invasive, low-cost nasal balloon autoinflation device has been found beneficial in other populations but has not been investigated among Aboriginal and Torres Strait Islander children. METHODS/DESIGN: This multi-centre, open-label, randomised controlled trial will determine the effectiveness of nasal balloon autoinflation compared to no nasal balloon autoinflation, for the treatment of OME among Aboriginal and Torres Strait Islander children in Australia. Children aged 3-16 years with unilateral or bilateral OME are being recruited from Aboriginal Health Services and the community. The primary outcome is the proportion of children showing tympanometric improvement of OME at 1 month. Improvement is defined as a change from bilateral type B tympanograms to at least one type A or C1 tympanogram, or from unilateral type B tympanogram to type A or C1 tympanogram in the index ear, without deterioration (type A or C1 to type C2, C3, or B tympanogram) in the contralateral ear. A sample size of 340 children (170 in each group) at 1 month will detect an absolute difference of 15% between groups with 80% power at 5% significance. Anticipating a 15% loss to follow-up, 400 children will be randomised. The primary analysis will be by intention to treat. Secondary outcomes include tympanometric changes at 3 and 6 months, hearing at 3 months, ear health-related quality of life (OMQ-14), and cost-effectiveness. A process evaluation including perspectives of parents or carers, health care providers, and researchers on trial implementation will also be undertaken. DISCUSSION: INFLATE will answer the important clinical question of whether nasal balloon autoinflation is an effective and acceptable treatment for Aboriginal and Torres Strait Islander children with OME. INFLATE will help fill the evidence gap for safe, low-cost, accessible OME therapies. TRIAL REGISTRATION: Australia New Zealand Clinical Trials Registry ACTRN12617001652369 . Registered on 22 December 2017. The Australia New Zealand Clinical Trials Registry is a primary registry of the WHO ICTRP network and includes all items from the WHO Trial Registration data set. Retrospective registration.

EISI: Extended inter-spike interval for mental health patients clustering based on mental health services and medications utilisation.

Mental health is vital in all human life stages, and managing mental healthcare service resources is crucial for providers. This paper presents a new method, called Extended Inter-Spike Interval (EISI), on identifying the patients with a similar utilisation of mental health services and medications. The EISI measures the distance between the utilisation patterns of the patients. Then, the pairwise distances are given to a developed split-and-merge Partitioning Around Medoids (PAM) clustering algorithm to identify the patients with similar utilisation patterns. To evaluate the proposed method, we use two years (2013-2014) of the 10% publicly available sample of the Australian Medicare Benefits Schedule (MBS) and Pharmaceutical Benefits Scheme (PBS) administrative data. Results show that mental health patients can be grouped into ten clusters with distinct and interpretable utilisations patterns. The largest cluster comprises individuals who only visit general practitioners and take psycholeptics medications for a short time. The smallest group contains occasional visits with general practitioners and regularly utilises psycholeptics and psychoanaleptics medications over long periods. The proposed method provides insights on whom to target and how to structure services for different groups of individuals with mental health conditions.

Do Financial Incentives Affect Utilization for Chronically Ill Medicare Beneficiaries?

OBJECTIVE: The objective of this study was to examine the price sensitivity for provider visits among Medicare Advantage beneficiaries. DATA SOURCES: We used Medicare Advantage encounter data from 2014 to 2017 accessed as part of an evaluation for the Center for Medicare & Medicaid Innovation. STUDY DESIGN: We analyzed the effect of cost-sharing on the utilization of 2 outcome categories: number of visits (specialist and primary care) and the probability of any visit (specialist and primary care). Our main independent variable was the size of the copayment for the visit, which we regressed on the outcomes with several beneficiary-level and plan-level control variables. DATA COLLECTION/EXTRACTION METHODS: We included beneficiaries with at least 1 of 4 specific chronic conditions and matched comparison beneficiaries. We did not require beneficiaries to be continuously enrolled from 2014 to 2017, but we required a full year of data for each year they were observed. This resulted in 371,140 beneficiary-year observations. PRINCIPAL FINDINGS: Copay reductions were associated with increases in utilization, although the changes were small, with elasticities <-0.2. We also found evidence of substitution effects between primary care provider (PCP) and specialist visits, particularly cardiology and endocrinology. When PCP copays declined, visits to these specialists also declined. CONCLUSIONS: We find that individuals with chronic conditions respond to changes in copays, although these responses are small. Reductions in PCP copays lead to reduced use of some specialists, suggesting that lowering PCP copays could be an effective way to reduce the use of specialist care, a desirable outcome if specialists are overused.

Bayesian prediction of emergency department wait time.

Increasingly, many hospitals are attempting to provide more accurate information about Emergency Department (ED) wait time to their patients. Estimation of ED wait time usually depends on what is known about the patient and also the status of the ED at the time of presentation. We provide a model for estimating ED wait time for prospective low acuity patients accessing information online prior to arrival. Little is known about the prospective patient and their condition. We develop a Bayesian quantile regression approach to provide an estimated wait time range for prospective patients. Our proposed approach incorporates a priori information in government statistics and elicited expert opinion. This methodology is compared to frequentist quantile regression and Bayesian quantile regression with non-informative priors. The test set includes 1, 024 low acuity presentations, of which 457 (44%) are Category 3, 425 (41%) are Category 4 and 160 (15%) are Category 5. On the Huber loss metric, the proposed method performs best on the test data for both median and 90th percentile prediction compared to non-informative Bayesian quantile regression and frequentist quantile regression. We obtain a benefit in the estimation of model coefficients due to the value contributed by a priori information in the form of elicited expert guesses guided by government wait time statistics. The use of such informative priors offers a beneficial approach to ED wait time prediction with demonstrable potential to improve wait time quantile estimates.

Artificial intelligence applications used in the clinical response to COVID-19: A scoping review.

Research into using artificial intelligence (AI) in health care is growing and several observers predicted that AI would play a key role in the clinical response to the COVID-19. Many AI models have been proposed though previous reviews have identified only a few applications used in clinical practice. In this study, we aim to (1) identify and characterize AI applications used in the clinical response to COVID-19; (2) examine the timing, location, and extent of their use; (3) examine how they relate to pre-pandemic applications and the U.S. regulatory approval process; and (4) characterize the evidence that is available to support their use. We searched academic and grey literature sources to identify 66 AI applications that performed a wide range of diagnostic, prognostic, and triage functions in the clinical response to COVID-19. Many were deployed early in the pandemic and most were used in the U.S., other high-income countries, or China. While some applications were used to care for hundreds of thousands of patients, others were used to an unknown or limited extent. We found studies supporting the use of 39 applications, though few of these were independent evaluations and we found no clinical trials evaluating any application's impact on patient health. Due to limited evidence, it is impossible to determine the extent to which the clinical use of AI in the pandemic response has benefited patients overall. Further research is needed, particularly independent evaluations on AI application performance and health impacts in real-world care settings.

Diabetes and the use of primary care provider services in rural, remote and metropolitan Australia.

INTRODUCTION: Public health agencies around the world are concerned about an ever-increasing burden of type 2 diabetes and related disability. Access to primary care providers (PCPs) can support early diagnosis and management. However, there is limited literature on how frequently older people with diabetes access PCPs, and their levels of access in rural Australia relative to metropolitan areas. METHODS: In this research, patterns of PCP use among those with diagnosed diabetes and those without diagnosed diabetes (referred to as 'healthy' individuals) were compared using a large survey of more than 230 000 people aged 45 years and older from New South Wales, Australia. A published model to study the PCP access patterns of a group of individuals with diabetes risk was used. RESULTS: Annual visits to PCPs among people aged 45 years or more with diabetes in rural areas, while higher than for healthy rural residents, were significantly lower than their metropolitan counterparts, mirroring similar disparities in PCP use across the rural-urban divide in the healthy population. Similar patterns were present in the high-risk population. Nevertheless, people with diabetes visited PCPs around four times a year, which is around the recommended number of annual visits, although some groups (eg those with comorbidities) may need more visits. CONCLUSION: Patterns of PCP use among rural residents, while significantly less frequent than their metropolitan counterparts, are at the recommended level for people with diabetes.

The economic cost of bulimia nervosa and binge eating disorder: a population-based study.

BACKGROUND: The most common eating disorders (EDs) are bulimia nervosa (BN) and binge eating disorder (BED), serious psychiatric illnesses that have devastating effects on the physical and psychological wellbeing of sufferers. EDs range in complexity and severity but can be life-threatening without appropriate treatment. Although it is well-known that quality of life impacts is high for ED sufferers, research regarding fiscal and related costs is severely limited. The aim of this study was to understand economic and other costs of EDs at the community level. METHOD: Data were derived from 2017 household community representative structured interview of 2977 people aged ⩾ 15 years in South Australia. ED diagnoses, health systems, productivity, transaction, out-of-pocket expenses and other related costs of BN and BED were used to estimate the economic burden of EDs in South Australia. RESULTS: The annual total economic cost of EDs in 2018 was estimated at $84 billion for South Australia. This included $81 billion from the burden of disease as the result of years lived with disability (YLD) ($62 billion) and years of life lost ($19 billion). The health system costs, productivity and tax revenue loss to the Australian economy were estimated at $1 billion, $1.6 billion and $0.6 billion, respectively. CONCLUSIONS: The YLD average cost in 2018 in South Australia was $296 649 per person. This is two-thirds of the costs borne by individuals and the wider economy. Prevention and management initiatives for EDs need to take into account these costs when assessing their potential benefits.

Multichannel mixture models for time-series analysis and classification of engagement with multiple health services: An application to psychology and physiotherapy utilization patterns after traffic accidents.

BACKGROUND: Motor vehicle accidents (MVA) represent a significant burden on health systems globally. Tens of thousands of people are injured in Australia every year and may experience significant disability. Associated economic costs are substantial. There is little literature on the health service utilization patterns of MVA patients. To fill this gap, this study has been designed to investigate temporal patterns of psychology and physiotherapy service utilization following transport-related injuries. METHOD: De-identified compensation data was provided by the Australian Transport Accident Commission. Utilization of physiotherapy and psychology services was analysed. The datasets contained 788 psychology and 3115 physiotherapy claimants and 22,522 and 118,453 episodes of service utilization, respectively. 582 claimants used both services, and their data were preprocessed to generate multidimensional time series. Time series clustering was applied using a mixture of hidden Markov models to identify the main distinct patterns of service utilization. Combinations of hidden states and clusters were evaluated and optimized using the Bayesian information criterion and interpretability. Cluster membership was further investigated using static covariates and multinomial logistic regression, and classified using high-performing classifiers (extreme gradient boosting machine, random forest and support vector machine) with 5-fold cross-validation. RESULTS: Four clusters of claimants were obtained from the clustering of the time series of service utilization. Service volumes and costs increased progressively from clusters 1 to 4. Membership of cluster 1 was positively associated with nerve damage and negatively associated with severe ABI and spinal injuries. Cluster 3 was positively associated with severe ABI, brain/head injury and psychiatric injury. Cluster 4 was positively associated with internal injuries. The classifiers were capable of classifying cluster membership with moderate to strong performance (AUC: 0.62-0.96). CONCLUSION: The available time series of post-accident psychology and physiotherapy service utilization were coalesced into four clusters that were clearly distinct in terms of patterns of utilization. In addition, pre-treatment covariates allowed prediction of a claimant's post-accident service utilization with reasonable accuracy. Such results can be useful for a range of decision-making processes, including the design of interventions aimed at improving claimant care and recovery.

The AEDUCATE Collaboration. Comprehensive antenatal education birth preparation programmes to reduce the rates of caesarean section in nulliparous women. Protocol for an individual participant data prospective meta-analysis.

INTRODUCTION: Rates of medical interventions in normal labour and birth are increasing. This prospective meta-analysis (PMA) proposes to assess whether the addition of a comprehensive multicomponent birth preparation programme reduces caesarean section (CS) in nulliparous women compared with standard hospital care. Additionally, do participant characteristics, intervention components or hospital characteristics modify the effectiveness of the programme? METHODS AND ANALYSIS: Population: women with singleton vertex pregnancies, no planned caesarean section (CS) or epidural.Intervention: in addition to hospital-based standard care, a comprehensive antenatal education programme that includes multiple components for birth preparation, addressing the three objectives: preparing women and their birth partner/support person for childbirth through education on physiological/hormonal birth (knowledge and understanding); building women's confidence through psychological preparation (positive mindset) and support their ability to birth without pain relief using evidence-based tools (tools and techniques). The intervention could occur in a hospital-based or community setting.Comparator: standard care alone in hospital-based maternity units. OUTCOMES: Primary: CS.Secondary: epidural analgesia, mode of birth, perineal trauma, postpartum haemorrhage, newborn resuscitation, psychosocial well-being.Subgroup analysis: parity, model of care, maternal risk status, maternal education, maternal socio-economic status, intervention components. STUDY DESIGN: An individual participant data (IPD) prospective meta-analysis (PMA) of randomised controlled trials, including cluster design. Each trial is conducted independently but share core protocol elements to contribute data to the PMA. Participating trials are deemed eligible for the PMA if their results are not yet known outside their Data Monitoring Committees. ETHICS AND DISSEMINATION: Participants in the individual trials will consent to participation, with respective trials receiving ethical approval by their local Human Research Ethics Committees. Individual datasets remain the property of trialists, and can be published prior to the publication of final PMA results. The overall data for meta-analysis will be held, analysed and published by the collaborative group, led by the Cochrane PMA group. TRIAL REGISTRATION NUMBER: CRD42020103857.

Cohort Profile: Effectiveness of a 12-Month Patient-Centred Medical Home Model Versus Standard Care for Chronic Disease Management among Primary Care Patients in Sydney, Australia.

Evidence suggests that patient-centred medical home (PCMH) is more effective than standard general practitioner care in improving patient outcomes in primary care. This paper reports on the design, early implementation experiences, and early findings of the 12-month PCMH model called 'WellNet' delivered across six primary care practices in Sydney, Australia. The WellNet study sample comprises 589 consented participants in the intervention group receiving enhanced primary care in the form of patient-tailored chronic disease management plan, improved self-management support, and regular monitoring by general practitioners (GPs) and trained clinical coordinators. The comparison group consisted of 7750 patients who were matched based on age, gender, type and number of chronic diseases who received standard GP care. Data collected include sociodemographic characteristics, clinical measures, and self-reported health assessments at baseline and 12 months. Early study findings show the mean age of the study participants was 70 years with nearly even gender distribution of males (49.7%) and females (50.3%). The most prevalent chronic diseases in descending order were circulatory system disorders (69.8%), diabetes (47.4%), musculoskeletal disorders (43.5%), respiratory diseases (28.7%), mental illness (18.8%), and cancer (13.6%). To our knowledge, the WellNet study is the first study in Australia to generate evidence on the feasibility of design, recruitment, and implementation of a comprehensive PCMH model. Lessons learned from WellNet study may inform other medical home models in Australian primary care settings.

Incorporation of expert knowledge in the statistical detection of diagnosis related group misclassification.

BACKGROUND: In activity based funding systems, the misclassification of inpatient episode Diagnostic Related Groups (DRGs) can have significant impacts on the revenue of health care providers. Weakly informative Bayesian models can be used to estimate an episode's probability of DRG misclassification. METHODS: This study proposes a new, Hybrid prior approach which utilises guesses that are elicited from a clinical coding auditor, switching to non-informative priors where this information is inadequate. This model's ability to detect DRG revision is compared to benchmark weakly informative Bayesian models and maximum likelihood estimates. RESULTS: Based on repeated 5-fold cross-validation, classification performance was greatest for the Hybrid prior model, which achieved best classification accuracy in 14 out of 20 trials, significantly outperforming benchmark models. CONCLUSIONS: The incorporation of elicited expert guesses via a Hybrid prior produced a significant improvement in DRG error detection; hence, it has the ability to enhance the efficiency of clinical coding audits when put into practice at a health care provider.

Predicting diabetes second-line therapy initiation in the Australian population via time span-guided neural attention network.

INTRODUCTION: The first line of treatment for people with Diabetes mellitus is metformin. However, over the course of the disease metformin may fail to achieve appropriate glycemic control, and a second-line therapy may become necessary. In this paper we introduce Tangle, a time span-guided neural attention model that can accurately and timely predict the upcoming need for a second-line diabetes therapy from administrative data in the Australian adult population. The method is suitable for designing automatic therapy review recommendations for patients and their providers without the need to collect clinical measures. DATA: We analyzed seven years of de-identified records (2008-2014) of the 10% publicly available linked sample of Medicare Benefits Schedule (MBS) and Pharmaceutical Benefits Scheme (PBS) electronic databases of Australia. METHODS: By design, Tangle inherits the representational power of pre-trained word embedding, such as GloVe, to encode sequences of claims with the related MBS codes. Moreover, the proposed attention mechanism natively exploits the information hidden in the time span between two successive claims (measured in number of days). We compared the proposed method against state-of-the-art sequence classification methods. RESULTS: Tangle outperforms state-of-the-art recurrent neural networks, including attention-based models. In particular, when the proposed time span-guided attention strategy is coupled with pre-trained embedding methods, the model performance reaches an Area Under the ROC Curve of 90%, an improvement of almost 10 percentage points over an attentionless recurrent architecture. IMPLEMENTATION: Tangle is implemented in Python using Keras and it is hosted on GitHub at https://github.com/samuelefiorini/tangle.

The association between physical activity and hospital payments for acute admissions in the Australian population aged 45 and over.

Physical activity (PA) is a key component of a healthy life, and it is hypothesised that individuals with higher levels of PA utilise fewer hospital resources. Quantifying the association between PA and hospital resource use is of interest to both payers and planners but estimates of its size in the general population are rare. In this paper we provide estimates of the association between PA and payments to hospitals in the Australian population over age 45. We use data from 45 and Up Study, a survey that contains health and lifestyle factors information about approximately 260,000 individuals over age 45 living in NSW, linked to hospital and death data. The linked data set allows to define a unique indicator for the level of PA over the week prior to the survey interview and to calculate payments to hospitals over the next year. We use Coarsened Exact Matching and multivariate analysis to study the relationship between PA and hospital payments, controlling for chronic health conditions, risk factors, standard socioeconomic variables and death. Our results clearly indicate that there is a statistically significant association between PA and lower hospital payments. While the size of the association depends to some extent on the covariates used in the model the conclusions are robust to changes in model specification. We also perform a sub-group analysis and show that the cost savings associated with PA are significantly larger for older and lower income populations. This study shows that if one is interested in lowering hospital expenditures then increasing PA levels is a policy that has the potential of being effective. It also shows that one does not need to target the entire population to achieve cost savings but can limit the intervention to the older population and/or the one in the lowest socioeconomic status.

Correction: Analysis of healthcare service utilization after transport-related injuries by a mixture of hidden Markov models.

[This corrects the article DOI: 10.1371/journal.pone.0206274.].