Machine-Learning Algorithms to Code Public Health Spending Accounts.

OBJECTIVES: Government public health expenditure data sets require time- and labor-intensive manipulation to summarize results that public health policy makers can use. Our objective was to compare the performances of machine-learning algorithms with manual classification of public health expenditures to determine if machines could provide a faster, cheaper alternative to manual classification. METHODS: We used machine-learning algorithms to replicate the process of manually classifying state public health expenditures, using the standardized public health spending categories from the Foundational Public Health Services model and a large data set from the US Census Bureau. We obtained a data set of 1.9 million individual expenditure items from 2000 to 2013. We collapsed these data into 147 280 summary expenditure records, and we followed a standardized method of manually classifying each expenditure record as public health, maybe public health, or not public health. We then trained 9 machine-learning algorithms to replicate the manual process. We calculated recall, precision, and coverage rates to measure the performance of individual and ensembled algorithms. RESULTS: Compared with manual classification, the machine-learning random forests algorithm produced 84% recall and 91% precision. With algorithm ensembling, we achieved our target criterion of 90% recall by using a consensus ensemble of ≥6 algorithms while still retaining 93% coverage, leaving only 7% of the summary expenditure records unclassified. CONCLUSIONS: Machine learning can be a time- and cost-saving tool for estimating public health spending in the United States. It can be used with standardized public health spending categories based on the Foundational Public Health Services model to help parse public health expenditure information from other types of health-related spending, provide data that are more comparable across public health organizations, and evaluate the impact of evidence-based public health resource allocation.

Detecting Proximal Secondary Caries Lesions: A Cost-effectiveness Analysis.

When choosing detection methods for secondary caries lesions, dentists need to weigh sensitivity, allowing early initiation of retreatments to avoid lesion progression, against specificity, aiming to reduce risks of false-positive diagnoses and invasive overtreatments. We assessed the cost-effectiveness of different detection methods for proximal secondary lesions using Monte Carlo microsimulations. A vital permanent molar with an occlusal-proximal restoration was simulated over the lifetime of an initially 20-y-old. Three methods were compared: biannual tactile detection, radiographic detection every 2 y, and biannual laser fluorescence detection. Methods were employed either on their own or in pairwise combinations at sensitive and specific thresholds estimated with systematically collected data. A mixed public-private payer perspective in the context of German health care was applied. Effectiveness was calculated as years of tooth retention. Net-benefit analyses were used to evaluate cost-effectiveness acceptability at different willingness-to-pay thresholds. Radiographic detection verified by tactile assessment (both at specific thresholds) was least costly (mean, 1,060 euros) but had limited effectiveness (mean retention time, 50 y). The most effective but also more costly combination was laser fluorescence detection verified by radiography, again at specific thresholds (1157 euros, 53 y, acceptable if willingness to pay >32 euro/y). In the majority of simulations, not combining detection methods or applying them at sensitive thresholds was less effective and more costly. Net benefits were not greatly altered by applying different discounting rates or using different baseline prevalence of secondary lesions. Current detection methods for secondary lesions should best be used in combination, not on their own, at specific thresholds to avoid false-positive diagnoses leading to costly and invasive overtreatment. The relevant characteristics, such as predictive value, of different methods should be assessed in longitudinal clinical studies.

Transaction costs of access to health care: Implications of the care-seeking pathways of tuberculosis patients for health system governance in Nigeria.

Health care costs incurred prior to the appropriate patient-provider transaction (i.e., transaction costs of access to health care) are potential barriers to accessing health care in low- and middle-income countries. This paper explores these transaction costs and their implications for health system governance through a cross-sectional survey of adult patients who received their first diagnosis of pulmonary tuberculosis (TB) at the three designated secondary health centres for TB care in Ebonyi State, Nigeria. The patients provided information on their care-seeking pathways and the associated costs prior to reaching the appropriate provider. Of the 452 patients, 84% first consulted an inappropriate provider. Only 33% of inappropriate consultations were with qualified providers (QP); the rest were with informal providers such as pharmacy providers (PPs; 57%) and traditional providers (TP; 10%). Notably, 62% of total transaction costs were incurred during the first visit to an inappropriate provider and the mean transaction costs incurred was highest with QPs (US$30.20) compared with PPs (US$14.40) and TPs (US$15.70). These suggest that interventions for reducing transaction costs should include effective decentralisation to integrate TB care with services at the primary health care level, community engagement to address information asymmetry, enforcing regulations to keep informal providers within legal limits and facilitating referral linkages among formal and informal providers to increase early contact with appropriate providers.

Decomposition of the drivers of the U.S. hospital spending growth, 2001-2009.

BACKGROUND: United States health care spending rose rapidly in the 2000s, after a period of temporary slowdown in the 1990s. However, the description of the overall trend and the understanding of the underlying drivers of this trend are very limited. This study investigates how well historical hospital cost/revenue drivers explain the recent hospital spending trend in the 2000s, and how important each of these drivers is. METHODS: We used aggregated time series data to describe the trend in total hospital spending, price, and quantity between 2001 and 2009. We used the Oaxaca-Blinder method to investigate the relative importance of major hospital cost/spending drivers (derived from the literature) in explaining the change in hospital spending patterns between 2001 and 2007. We assembled data from Medicare Cost Reports, American Hospital Association annual surveys, Prospective Payment System (PPS) Impact Files, Medicare Provider Analysis and Review (MedPAR) Medicare claims data, InterStudy reports, National Health Expenditure data, and Area Resource Files. RESULTS: Aggregated time series trends show that high hospital spending between 2001 and 2009 appears to be driven by higher payment per unit of hospital output, not by increased utilization. Results using the Oaxaca-Blinder regression decomposition method indicate that changes in historically important spending drivers explain a limited 30% of unit-payment growth, but a higher 60% of utilization growth. Hospital staffing and labor-related costs, casemix, and demographics are the most important drivers of higher hospital revenue, utilization, and unit-payment. Technology is associated with lower utilization, higher unit payment, and limited increases in total revenue. Market competition, primarily because of increased managed care concentration, moderates total revenue growth by driving lower unit payment. CONCLUSIONS: Much of the rapidly rising hospital spending growth in the 2000s in the United States is driven by factors not commonly known or well measured. Future studies need to explore new factors and dynamics that drive longer-term hospital spending growth in recent years, particularly through the channel of higher prices.

Health issues and health care expenses in Canadian bankruptcies and insolvencies.

Illness can contribute to financial problems directly, through high medical bills, and indirectly, through lost income. No previous in-depth studies have documented the role of medical problems among Canadian bankruptcy filers. We obtained the bankruptcy filings from a random sample of 5,000 debtors across Canada and mailed surveys to them seeking information about the medical antecedents of their bankruptcy. A total of 521 debtors responded (response rate of 10.4%), of whom 40.1 percent reported losing at least two weeks of work-related income because of illness or injury in the two years before their filing; 8.3 percent reported a similar income loss because of caregiving responsibilities for someone else who was ill. Although 60.1 percent of respondents reported being responsible for a medical bill within the previous two years, only 6.9 percent had bills over $5,000 (all amounts in Canadian Dollars). Prescription drugs were cited as the costliest medical expense by two-thirds of debtors reporting bills > $5,000, with dental bills cited by 22.2 percent. Universal health insurance affords Canadians protection against ruinous doctor and hospital bills. Inadequate coverage for prescription drugs and dental care, however, leaves some with unaffordable out-of-pocket costs. In addition, illness is a frequent indirect cause of bankruptcy through loss of work-related income.

Orthopedics coding and funding.

The French tarification à l'activité (T2A) prospective payment system is a financial system in which a health-care institution's resources are based on performed activity. Activity is described via the PMSI medical information system (programme de médicalisation du système d'information). The PMSI classifies hospital cases by clinical and economic categories known as diagnosis-related groups (DRG), each with an associated price tag. Coding a hospital case involves giving as realistic a description as possible so as to categorize it in the right DRG and thus ensure appropriate payment. For this, it is essential to understand what determines the pricing of inpatient stay: namely, the code for the surgical procedure, the patient's principal diagnosis (reason for admission), codes for comorbidities (everything that adds to management burden), and the management of the length of inpatient stay. The PMSI is used to analyze the institution's activity and dynamism: change on previous year, relation to target, and comparison with competing institutions based on indicators such as the mean length of stay performance indicator (MLS PI). The T2A system improves overall care efficiency. Quality of care, however, is not presently taken account of in the payment made to the institution, as there are no indicators for this; work needs to be done on this topic.

Medical expenditures associated with nonfatal occupational injuries among immigrant and U.S.-born workers.

BACKGROUND: No national study has investigated whether immigrant workers are less likely than U.S.-workers to seek medical treatment after occupational injuries and whether the payment source differs between two groups. METHODS: Using the 2004-2009 Medical Expenditure Panel Survey (MEPS) data, we estimated the annual incidence rate of nonfatal occupational injuries per 100 workers. Logistic regression models were fitted to test whether injured immigrant workers were less likely than U.S.-born workers to seek professional medical treatment after occupational injuries. We also estimated the average mean medical expenditures per injured worker during the 2 year MEPS reference period using linear regression analysis, adjusting for gender, age, race, marital status, education, poverty level, and insurance. Types of service and sources of payment were compared between U.S.-born and immigrant workers. RESULTS: A total of 1,909 injured U.S.-born workers reported 2,176 occupational injury events and 508 injured immigrant workers reported 560 occupational injury events. The annual nonfatal incidence rate per 100 workers was 4.0% (95% CI: 3.8%-4.3%) for U.S.-born workers and 3.0% (95% CI: 2.6%-3.3%) for immigrant workers. Medical treatment was sought after 77.3% (95% CI: 75.1%-79.4%) of the occupational injuries suffered by U.S.-born workers and 75.6% (95% CI: 69.8%-80.7%) of the occupational injuries suffered by immigrant workers. The average medical expenditure per injured worker in the 2 year MEPS reference period was $2357 for the U.S.-born workers and $2,351 for immigrant workers (in 2009 U.S. dollars, P = 0.99). Workers' compensation paid 57.0% (95% CI: 49.4%-63.6%) of the total expenditures for U.S.-born workers and 43.2% (95% CI: 33.0%-53.7%) for immigrant workers. U.S.-born workers paid 6.7% (95% CI: 5.5%-8.3%) and immigrant workers paid 7.1% (95% CI: 5.2%-9.6%) out-of-pocket. CONCLUSIONS: Immigrant workers had a statistically significant lower incidence rate of nonfatal occupational injuries than U.S.-born workers. There was no significant difference in seeking medical treatment and in the mean expenditures per injured worker between the two groups. The proportion of total expenditures paid by workers' compensation was smaller (marginally significant) for immigrant workers than for U.S.-born workers.

Autism spectrum disorders and health care expenditures: the effects of co-occurring conditions.

OBJECTIVE: Children with autism spectrum disorders (ASDs) often have co-occurring conditions, but little is known on the effect of those conditions on their medical care cost. Medical expenditures attributable to ASDs among Medicaid-enrolled children were calculated, and the effects of 3 commonly co-occurring conditions--intellectual disability (ID), attention deficit/hyperactivity disorder (ADHD), and epilepsy-on those expenditures were analyzed. METHODS: Using MarketScan Medicaid Multi-State Databases (2003-2005) and the International Classification of Disease, Ninth Revision, children with ASD were identified. Children without ASD formed the comparison group. The 3 co-occurring conditions were identified among both the ASD and the comparison groups. Annual mean, median, and 95th percentile of total expenditures were calculated for children with ASD and the co-occurring conditions and compared with those of children without ASD. Multivariate analyses established the influence of each of those co-occurring conditions on the average expenditures for children with and without ASD. RESULTS: In 2005, 47% of children with ASD had at least 1 selected co-occurring condition; attention deficit/hyperactivity disorder was the most common, at 30%. The mean medical expenditures for children with ASD were 6 times higher than those of the comparison group. Children with ASD and ID incurred expenditures 2.7 times higher than did children with ASD and no co-occurring condition. CONCLUSION: Medicaid-enrolled children with ASD incurred higher medical costs than did Medicaid-enrolled children without ASD. Among Medicaid-enrolled children with ASD, cost varied substantially based on the presence of another neurodevelopmental disorder. In particular, children with ID had much higher costs than did other children with ASD.

La priorización de fármacos oncológicos en el sistema hospitalario de Cataluña: estudio cualitativo de casos / Prioritization of chemotherapy drugs in the Catalan hospital system: a qualitative case study

Objetivos Analizar el proceso de priorización de fármacos oncológicos en las comisiones farmacoterapéuticas (CFT) de los hospitales de Cataluña y examinar el grado en que influyen en el acceso de los pacientes a estos fármacos.Métodos Estudio cualitativo de casos de las CFT de hospitales de tercer nivel de Cataluña basado en entrevistas semiestructuradas y en una revisión de la literatura científica. Los sujetos de estudio son profesionales que pueden aportar una visión técnica del funcionamiento de las CFT, entre ellos farmacéuticos, oncólogos médicos, farmacólogos clínicos y profesionales de otras especialidades médicas, y otra visión de carácter institucional sobre el marco de gestión hospitalario y autonómico. Para el análisis del proceso de establecimiento de prioridades se ha utilizado el marco conceptual conocido como «justificación de la acción responsable», de Daniels y Sabin, que propone una perspectiva analítica de la toma de decisiones justas y legítimas. Resultados El estudio permite identificar las debilidades del actual marco regulador en la introducción de fármacos, fragmentado por hospital y carente de estrategias de coordinación que permitan priorizar y optimizar recursos en el conjunto del sistema sanitario catalán. Conclusión Se propone desarrollar una estrategia de coordinación de las decisiones para todo el sector público hospitalario con el fin de afrontar un entorno cada vez más innovador en el cual se eviten las desigualdades de acceso (AU)

Objectives To analyze the prioritization process for chemotherapy drugs in the Drug-Therapeutic Committees (DTCs) in Catalan hospitals and assess their impact on patients’ access to these drugs.Methods A case qualitative study of the DTCs of tertiary hospitals in Catalonia was performed, based on semi-structured interviews and a review of the scientific literature. Key professionals were interviewed with technical and institutional involvement in the DTCs assessment process, including pharmacists, oncologists, clinical pharmacologists and other physicians with different medical specialties. The conceptual framework "accountability for reasonableness", by Daniels and Sabin, which analyzes prioritization processes from a “fair and legitimate decision-making” perspective, was used.Results The current regulatory framework for the introduction of drugs is characterized by a fragmented hospital system and lacks coordination strategies. These weaknesses fail to allow the prioritization and optimization of resources in the Catalan health service.Conclusion We propose the implementation of a decision-making coordination strategy for all public sector hospitals. This strategy would allow an increasingly innovative environment to be exploited, in which inequalities in access to drugs at the hospital level should be avoided (AU)

Reducing out-of-pocket expenditures to reduce poverty: a disaggregated analysis at rural-urban and state level in India.

Out-of-pocket (OOP) expenditure on health care has significant implications for poverty in many developing countries. This paper aims to assess the differential impact of OOP expenditure and its components, such as expenditure on inpatient care, outpatient care and on drugs, across different income quintiles, between developed and less developed regions in India. It also attempts to measure poverty at disaggregated rural-urban and state levels. Based on Consumer Expenditure Survey (CES) data from the National Sample Survey (NSS), conducted in 1999-2000, the share of households' expenditure on health services and drugs was calculated. The number of individuals below the state-specific rural and urban poverty line in 17 major states, with and without netting out OOP expenditure, was determined. This also enabled the calculation of the poverty gap or poverty deepening in each region. Estimates show that OOP expenditure is about 5% of total household expenditure (ranging from about 2% in Assam to almost 7% in Kerala) with a higher proportion being recorded in rural areas and affluent states. Purchase of drugs constitutes 70% of the total OOP expenditure. Approximately 32.5 million persons fell below the poverty line in 1999-2000 through OOP payments, implying that the overall poverty increase after accounting for OOP expenditure is 3.2% (as against a rise of 2.2% shown in earlier literature). Also, the poverty headcount increase and poverty deepening is much higher in poorer states and rural areas compared with affluent states and urban areas, except in the case of Maharashtra. High OOP payment share in total health expenditures did not always imply a high poverty headcount; state-specific economic and social factors played a role. The paper argues for better methods of capturing drugs expenditure in household surveys and recommends that special attention be paid to expenditures on drugs, in particular for the poor. Targeted policies in just five poor states to reduce OOP expenditure could help to prevent almost 60% of the poverty headcount increase through OOP payments.

Why U.S. health care expenditure and ranking on health care indicators are so different from Canada's.

Compared to other industrialized countries, the U.S. spends most of all on health care. Nonetheless, the U.S. ranks relatively low on health care indicators. This paradox has been already known for decades. For example, the turning point comparing the U.S. and Canada was in 1972. Health expenditure as a percentage of GDP was higher in Canada than in the USA from 1960 until 1972. Since 1972 expenditure on health care has been higher in the U.S. than in Canada (OECD 2005a, Health data 2005, fourteenth OECD electronic database on health systems, date of release June 2005, last update 04/26/2005). The present study integrates the dispersed literature on spending and health care rankings and adds some statistical analysis to these studies. The evaluation of different factors influencing health care expenditure in the U.S. relative to other countries is restricted to a comparison with Canada. The U.S. and Canada are two countries that are sufficiently similar to make comparisons useful. The comparison of factors influencing health care expenditure in the U.S. and Canada in 2002 reveals that health care expenditure in the U.S. is higher than in Canada mainly due to administration costs, Baumol's cost disease and pharmaceutical prices. It is not primarily inefficiency in health care production but the dominant prevalence for free choice and own responsibility that explains the paradox of high expenditure on health care and low ranking on health care indicators.

Private gain and public pain: financing American health care.

Health care spending comprises about 16% of the total United States gross domestic product and continues to rise. This article examines patterns of health care spending and the factors underlying their proportional growth. We examine the "usual suspects" most frequently cited as drivers of health care costs and explain why these may not be as important as they seem. We suggest that the drive for technological advancement, coupled with the entrepreneurial nature of the health care industry, has produced inherently inequitable and unsustainable health care expenditure and growth patterns. Successful health reform will need to address these factors and their consequences.