Project description:OBJECTIVE:To propose nonparametric double robust machine learning in variable importance analyses of medical conditions for health spending. DATA SOURCES:2011-2012 Truven MarketScan database. STUDY DESIGN:I evaluate how much more, on average, commercially insured enrollees with each of 26 of the most prevalent medical conditions cost per year after controlling for demographics and other medical conditions. This is accomplished within the nonparametric targeted learning framework, which incorporates ensemble machine learning. Previous literature studying the impact of medical conditions on health care spending has almost exclusively focused on parametric risk adjustment; thus, I compare my approach to parametric regression. PRINCIPAL FINDINGS:My results demonstrate that multiple sclerosis, congestive heart failure, severe cancers, major depression and bipolar disorders, and chronic hepatitis are the most costly medical conditions on average per individual. These findings differed from those obtained using parametric regression. CONCLUSIONS:The literature may be underestimating the spending contributions of several medical conditions, which is a potentially critical oversight. If current methods are not capturing the true incremental effect of medical conditions, undesirable incentives related to care may remain. Further work is needed to directly study these issues in the context of federal formulas.
Project description:ImportanceCareFirst, the largest commercial insurer in the mid-Atlantic Region of the United States, runs a medical home program focusing on financial incentives for primary care practices and care coordination for high-risk patients. From 2013 to 2015, CareFirst extended the program to Medicare fee-for-service (FFS) beneficiaries in participating practices. If the model extension improved quality while reducing spending, the Centers for Medicare and Medicaid Services could expand the program to Medicare beneficiaries broadly.ObjectiveTo test whether extending CareFirst's program to Medicare FFS patients improves care processes and reduces hospitalizations, emergency department visits, and spending.Design, setting, and participantsThis difference-in-differences analysis compared outcomes for roughly 35 000 Medicare FFS patients attributed to 52 intervention practices (grouped by CareFirst into 14 "medical panels") to outcomes for 69 000 Medicare patients attributed to 42 matched comparison panels during a 1-year baseline period and 2.5-year intervention at Maryland primary care practices.Main outcomes and measuresHospitalizations (all-cause and ambulatory-care sensitive), emergency department visits, Medicare Part A and B spending, and 3 quality-of-care process measures: ambulatory care within 14 days of a hospital stay, cholesterol testing for those with ischemic vascular disease, and a composite measure for those with diabetes.InterventionsCareFirst hired nurses who worked with patients' usual primary care practitioners to coordinate care for 3656 high-risk Medicare patients. CareFirst paid panels rewards for meeting cost and quality targets for their Medicare patients and advised panels on how to meet these targets based on analyses of claims data.ResultsOn average, each of the 14 intervention panels had 9.3 primary care practitioners and was attributed 2202 Medicare FFS patients in the baseline period. The panels' attributed Medicare patients were, on average, 73.8 years old, 59.2% female, and 85.1% white. The extension of CareFirst's program to Medicare patients was not statistically associated with improvements in any outcomes, either for the full Medicare population or for a high-risk subgroup in which impacts were expected to be largest. For the full population, the difference-in-differences estimates were 1.4 hospitalizations per 1000 patients per quarter (P = .54; 90% CI, -2.1 to 5.0), -2.5 outpatient ED visits per 1000 patients per quarter (P = .26; 90% CI, -6.2 to 1.1), and -$1 per patient per month in Medicare Part A and B spending (P = .98; 90% CI, -$40 to $39). For hospitalizations and Medicare spending, the 90% CIs did not span CareFirst's expected impacts. Hospitalizations for the intervention group declined by 10% from baseline year to the final 18 months of the intervention, but this was matched by similar declines in the comparison group.Conclusion and relevanceThe extension of CareFirst's program to Medicare did not measurably improve quality-of-care processes or reduce service use or spending for Medicare patients. Further program refinement and testing would be needed to support scaling the program more broadly to Medicare patients.
Project description:Many individuals visit rural telemedicine centres to obtain safe and effective health remedies for their physical and emotional illnesses. This study investigates the antecedents of patients' satisfaction relating to telemedicine adoption in rural public hospitals settings in Bangladesh through the adaptation of Expectation Disconfirmation Theory extended by Social Cognitive Theory. This research advances a theoretically sustained prediction model forecasting patients' satisfaction with telemedicine to enable informed decision making. A research model explores four potential antecedents: expectations, performance, disconfirmation, and enjoyment; that significantly contribute to predicting patients' satisfaction concerning telemedicine adoption in Bangladesh. This model is validated using two-staged structural equation modeling and artificial neural network approaches. The findings demonstrate the determinants of patients' satisfaction with telemedicine. The presented model will assist medical practitioners, academics, and information systems practitioners to develop high-quality decisions in the future application of telemedicine. Pertinent implications, limitations and future research directions are endorsed securing long-term telemedicine sustainability.
Project description:Given the complexity of high-acuity health care, designing an effective clinical note template can be beneficial to both document patient care and clarify how telemedicine is used. We characterized documented interactions via a standardized note template between bedside intensive care unit (ICU) providers and teleintensivists in 2 Veterans Health Administration ICU telemedicine support centers. All ICUs linked to support centers and providing care from October 2012 through September 2014 were considered. Interactions were assessed based on initiation site, bedside initiator, contact type, and patient care change. Of 14 511 ICU admissions with teleintensivist access, teleintensivist interaction was documented in 21.6% (N = 3136). In particular, contacts were primarily initiated by bedside staff (74.4%), use increased over time, and of contacts resulting in changes in patient care, most were initiated by a bedside nurse (84.3%). Given this variation, future research necessitates inclusion of utilization in evaluation of Tele-ICU and patient outcomes.
Project description:Early detection of bacteremia is important to prevent antibiotic abuse. Therefore, we aimed to develop a clinically applicable bacteremia prediction model using machine learning technology. Data from two tertiary medical centers' electronic medical records during a 12-year-period were extracted. Multi-layer perceptron (MLP), random forest, and gradient boosting algorithms were applied for machine learning analysis. Clinical data within 12 and 24 hours of blood culture were analyzed and compared. Out of 622,771 blood cultures, 38,752 episodes of bacteremia were identified. In MLP with 128 hidden layer nodes, the area under the receiver operating characteristic curve (AUROC) of the prediction performance in 12- and 24-h data models was 0.762 (95% confidence interval (CI); 0.7617-0.7623) and 0.753 (95% CI; 0.7520-0.7529), respectively. AUROC of causative-pathogen subgroup analysis predictive value for Acinetobacter baumannii bacteremia was the highest at 0.839 (95% CI; 0.8388-0.8394). Compared to primary bacteremia, AUROC of sepsis caused by pneumonia was highest. Predictive performance of bacteremia was superior in younger age groups. Bacteremia prediction using machine learning technology appeared possible for acute infectious diseases. This model was more suitable especially to pneumonia caused by Acinetobacter baumannii. From the 24-h blood culture data, bacteremia was predictable by substituting only the continuously variable values.
Project description:ImportanceKnowledge about use of health care services (health care utilization) and expenditures among unauthorized immigrant populations is uncertain because of limitations in ascertaining legal status in population data.ObjectiveTo examine health care utilization and expenditures that are attributable to unauthorized and authorized immigrants vs US-born individuals.Design, setting, and participantsThis cross-sectional study used the data on documentation status from the Los Angeles Family and Neighborhood Survey (LAFANS) to develop a random forest classifier machine learning model. K-fold cross-validation was used to test model performance. The LAFANS is a randomized, multilevel, in-person survey of households residing in Los Angeles County, California, consisting of 2 waves. Wave 1 began in April 2000 and ended in January 2002, and wave 2 began in August 2006 and ended in December 2008. The machine learning model was then applied to a nationally representative database, the 2016-2017 Medical Expenditure Panel Survey (MEPS), to predict health care expenditures and utilization among unauthorized and authorized immigrants and US-born individuals. A generalized linear model analyzed health care expenditures. Logistic regression modeling estimated dichotomous use of emergency department (ED), inpatient, outpatient, and office-based physician visits by immigrant groups with adjusting for confounding factors. Data were analyzed from May 1, 2019, to October 14, 2020.ExposuresSelf-reported immigration status (US-born, authorized, and unauthorized status).Main outcomes and measuresAnnual health care expenditures per capita and use of ED, outpatient, inpatient, and office-based physician care.ResultsOf 47 199 MEPS respondents with nonmissing data, 35 079 (74.3%) were US born, 10 816 (22.9%) were authorized immigrants, and 1304 (2.8%) were unauthorized immigrants (51.7% female; mean age, 47.6 [95% CI, 47.4-47.8] years). Compared with authorized immigrants and US-born individuals, unauthorized immigrants were more likely to be aged 18 to 44 years (80.8%), Latino (96.3%), and Spanish speaking (95.2%) and to have less than 12 years of education (53.7%). Half of unauthorized immigrants (47.1%) were uninsured compared with 15.9% of authorized immigrants and 6.0% of US-born individuals. Mean annual health care expenditures per person were $1629 (95% CI, $1330-$1928) for unauthorized immigrants, $3795 (95% CI, $3555-$4035) for authorized immigrants, and $6088 (95% CI, $5935-$6242) for US-born individuals.Conclusions and relevanceContrary to much political discourse in the US, this cross-sectional study found no evidence that unauthorized immigrants are a substantial economic burden on safety net facilities such as EDs. This study illustrates the value of machine learning in the study of unauthorized immigrants using large-scale, secondary databases.
Project description:BackgroundThis study examines the effects of a shift in medical coverage, from National Health Insurance (NHI) to Medical Aid (MA), on health care utilization (measured by the number of outpatient visits and length of stay; LOS) and out-of-pocket medical expenses.MethodsData were collected from the Korean Welfare Panel Study (2010-2016). A total of 888 MA Type I beneficiaries and 221 MA Type II beneficiaries who shifted from the NHI were included as the case group and 2664 and 663 consecutive NHI holders (1:3 propensity score-matched) were included as the control group, respectively. We used the 'difference-in-differences' (DiD) analysis approach to assess changes in health care utilization and medical spending by the group members.ResultsDifferential average changes in outpatient visits in the MA Type I panel between the pre- and post-shift periods were significant, but differential changes in LOS were not found. Those who shifted from NHI to MA Type I had increased number of outpatient visits without changes in out-of-pocket spending, compared to consecutive NHI holder who had similar characteristics. However, this was not found for MA Type II beneficiaries.ConclusionOur research provides evidence that the shift in medical coverage from NHI to MA Type I increased the number of outpatient visits without increasing the out-of-pocket spending. Considering the problem of excess medical utilization by Korean MA Type I beneficiaries, further researches are required to have in-depth discussions on the appropriateness of the current cost-sharing level on MA beneficiaries.
Project description:PurposeTo identify components of the patient-centered medical home (PCMH) model of care that are associated with lower spending and utilization among Medicare beneficiaries.MethodsRegression analyses of changes in outcomes for Medicare beneficiaries in practices that engaged in particular PCMH activities compared with beneficiaries in practices that did not. We analyzed claims for 302,719 Medicare fee-for-service beneficiaries linked to PCMH surveys completed by 394 practices in the Centers for Medicare & Medicaid Services' 8-state Multi-Payer Advanced Primary Care Practice demonstration.ResultsSix activities were associated with lower spending or utilization. Use of a registry to identify and remind patients due for preventive services was associated with all 4 of our outcome measures: total spending was $69.77 less per beneficiary per month (PBPM) (P = 0.00); acute-care hospital spending was $36.62 less PBPM (P = 0.00); there were 6.78 fewer hospital admissions per 1,000 beneficiaries per quarter (P1KBPQ) (P = 0.003); and 11.05 fewer emergency department (ED) visits P1KBPQ (P = 0.05). Using a patient registry for pre-visit planning and clinician reminders was associated with $29.31 lower total spending PBPM (P = 0.05). Engaging patients with chronic conditions in goal setting and action planning was associated with 4.62 fewer hospital admissions P1KBPQ (P = 0.01) and 11.53 fewer ED visits P1KBPQ (P = 0.00). Monitoring patients during hospital stays was associated with $22.06 lower hospital spending PBPM (P = 0.03). Developing referral protocols with commonly referred-to clinicians was associated with 11.62 fewer ED visits P1KBPQ (P = 0.00). Using quality improvement approaches was associated with 13.47 fewer ED visits P1KBPQ (P =0.00).ConclusionsPractices seeking to deliver more efficient care may benefit from implementing these 6 activities.
Project description:We compare healthcare spending in public and private Medicare using newly available claims data from Medicare Advantage (MA) insurers. MA insurer revenues are 30 percent higher than their healthcare spending. Adjusting for enrollee mix, healthcare spending per enrollee in MA is 9 to 30 percent lower than in traditional Medicare (TM), depending on the way we define "comparable" enrollees. Spending differences primarily reflect differences in healthcare utilization, with similar reductions for "high value" and "low value" care, rather than healthcare prices. We present evidence consistent with MA plans encouraging substitution to less expensive care and engaging in utilization management. (JEL H11, H42, H51, I11, I13).
Project description:Single molecule localisation (SML) microscopy is a fundamental tool for biological discoveries; it provides sub-diffraction spatial resolution images by detecting and localizing "all" the fluorescent molecules labeling the structure of interest. For this reason, the effective resolution of SML microscopy strictly depends on the algorithm used to detect and localize the single molecules from the series of microscopy frames. To adapt to the different imaging conditions that can occur in a SML experiment, all current localisation algorithms request, from the microscopy users, the choice of different parameters. This choice is not always easy and their wrong selection can lead to poor performance. Here we overcome this weakness with the use of machine learning. We propose a parameter-free pipeline for SML learning based on support vector machine (SVM). This strategy requires a short supervised training that consists in selecting by the user few fluorescent molecules (∼ 10-20) from the frames under analysis. The algorithm has been extensively tested on both synthetic and real acquisitions. Results are qualitatively and quantitatively consistent with the state of the art in SML microscopy and demonstrate that the introduction of machine learning can lead to a new class of algorithms competitive and conceived from the user point of view.