Machine learning methods are flexible prediction algorithms that may be more accurate than conventional regression. We compared the accuracy of different techniques for detecting clinical ...deterioration on the wards in a large, multicenter database.
Observational cohort study.
Five hospitals, from November 2008 until January 2013.
Hospitalized ward patients
None
Demographic variables, laboratory values, and vital signs were utilized in a discrete-time survival analysis framework to predict the combined outcome of cardiac arrest, intensive care unit transfer, or death. Two logistic regression models (one using linear predictor terms and a second utilizing restricted cubic splines) were compared to several different machine learning methods. The models were derived in the first 60% of the data by date and then validated in the next 40%. For model derivation, each event time window was matched to a non-event window. All models were compared to each other and to the Modified Early Warning score, a commonly cited early warning score, using the area under the receiver operating characteristic curve (AUC). A total of 269,999 patients were admitted, and 424 cardiac arrests, 13,188 intensive care unit transfers, and 2,840 deaths occurred in the study. In the validation dataset, the random forest model was the most accurate model (AUC, 0.80 95% CI, 0.80-0.80). The logistic regression model with spline predictors was more accurate than the model utilizing linear predictors (AUC, 0.77 vs 0.74; p < 0.01), and all models were more accurate than the MEWS (AUC, 0.70 95% CI, 0.70-0.70).
In this multicenter study, we found that several machine learning methods more accurately predicted clinical deterioration than logistic regression. Use of detection algorithms derived from these techniques may result in improved identification of critically ill patients on the wards.
Vitamin D treatment has been found to decrease the incidence of viral respiratory tract infection, especially in patients with vitamin D deficiency. Whether vitamin D is associated with coronavirus ...disease 2019 (COVID-19) incidence is unknown.
To examine whether the last vitamin D status before COVID-19 testing is associated with COVID-19 test results.
This retrospective cohort study at an urban academic medical center included patients with a 25-hydroxycholecalciferol or 1,25-dihydroxycholecalciferol level measured within 1 year before being tested for COVID-19 from March 3 to April 10, 2020.
Vitamin D deficiency was defined by the last measurement of 25-hydroxycholecalciferol less than 20 ng/mL or 1,25-dihydroxycholecalciferol less than 18 pg/mL before COVID-19 testing. Treatment changes were defined by changes in vitamin D type and dose between the date of the last vitamin D level measurement and the date of COVID-19 testing. Vitamin D deficiency and treatment changes were combined to categorize the most recent vitamin D status before COVID-19 testing as likely deficient (last level deficient and treatment not increased), likely sufficient (last level not deficient and treatment not decreased), and 2 groups with uncertain deficiency (last level deficient and treatment increased, and last level not deficient and treatment decreased).
The outcome was a positive COVID-19 polymerase chain reaction test result. Multivariable analysis tested whether vitamin D status before COVID-19 testing was associated with testing positive for COVID-19, controlling for demographic and comorbidity indicators.
A total of 489 patients (mean SD age, 49.2 18.4 years; 366 75% women; and 331 68% race other than White) had a vitamin D level measured in the year before COVID-19 testing. Vitamin D status before COVID-19 testing was categorized as likely deficient for 124 participants (25%), likely sufficient for 287 (59%), and uncertain for 78 (16%). Overall, 71 participants (15%) tested positive for COVID-19. In multivariate analysis, testing positive for COVID-19 was associated with increasing age up to age 50 years (relative risk, 1.06; 95% CI, 1.01-1.09; P = .02); non-White race (relative risk, 2.54; 95% CI, 1.26-5.12; P = .009), and likely deficient vitamin D status (relative risk, 1.77; 95% CI, 1.12-2.81; P = .02) compared with likely sufficient vitamin D status. Predicted COVID-19 rates in the deficient group were 21.6% (95% CI, 14.0%-29.2%) vs 12.2%(95% CI, 8.9%-15.4%) in the sufficient group.
In this single-center, retrospective cohort study, likely deficient vitamin D status was associated with increased COVID-19 risk, a finding that suggests that randomized trials may be needed to determine whether vitamin D affects COVID-19 risk.
To study cost-related medication non-adherence (CRN) for a 30-month period before and during the COVID-19 pandemic using a sample of Medicare patients at high risk of hospitalization.
A novel data ...set of quarterly surveys of CRN was used to evaluate CRN before and during the COVID-19 pandemic. Generalized Estimating Equation (GEE) analyses were conducted to evaluate the adjusted coefficients of change in CRN behaviors controlling for socio-demographic and health characteristics.
Six hundred seventy-seven Medicare beneficiaries at high risk of hospitalization who were alive on January 1, 2020 and followed up through quarterly surveys on CRN for 30 months before and during the COVID-19 pandemic.
Two metrics of prevalence and persistence of CRN and their adjusted coefficients in GEE with binomial family distribution and log link function controlling for socio-demographic and health characteristics.
A total of 5,990 quarterly surveys were completed by the 677 patients during the 30-month study period. Among the 677 patients, 250 (37%) were men, 591 (87%) were African American, and 288 (42%) were Medicare-Medicaid dual eligible. The unadjusted prevalence of CRN before and during the COVID-19 pandemic was 31.1% and 25.7% respectively (p = 0.02 by Chi-squared test), and persistent CRN rates were 12.1% and 9.7% respectively (p = 0.17 by Chi-squared test). The adjusted odds ratio of CRN prevalence during the pandemic compared to the pre-pandemic level was 0.75 (p<0.01), and 0.74 (p = 0.03) for persistent CRN in GEE estimations.
There are coherent evidence of a reversal of CRN rates during the COVID-19 pandemic among this high-need, high-cost resource utilization Medicare population. Patients' CRN behaviors may be responsive to exogenous impacts, and the behaviors changed in the same direction with similar magnitude in terms of prevalence (the extensive margin) and persistence (the intensive margin). More research is needed to advance the understanding of the driving forces behind patients' behavioral changes and to identify factors that may be informative for reducing CRN in the long run.
To explain prior literature showing that married Medicare beneficiaries achieve better health outcomes at half the per person cost of single beneficiaries, we examined different patterns of ...healthcare utilization as a potential driver.Using the Medicare Current Beneficiary Survey (MCBS) data, we sought to understand utilization patterns in married versus currently-not-married Medicare beneficiaries. We analyzed the relationship between marital status and healthcare utilization (classified based on setting of care utilization into outpatient, inpatient, and skilled nursing facility (SNF) use) using logistic regression modeling. We specified models to control for possible confounders based on the Andersen model of healthcare utilization.Based on 13,942 respondents in the MCBS dataset, 12,929 had complete data, thus forming the analytic sample, of whom 6473 (50.3%) were married. Of these, 58% (vs. 36% of those currently-not-married) were male, 45% (vs. 47%) were age >75, 24% (vs. 70%) had a household income below $25,000, 18% (vs. 14%) had excellent self-reported general health, and 56% (vs. 36%) had private insurance. Compared to unmarried respondents, married respondents had a trend toward higher odds of having a recent outpatient visit (unadjusted odds ratio (OR) 1.11, 95% confidence interval (CI) 1.04-1.19, adjusted odds ratio (AOR) 1.10, (CI) 0.99-1.22), and lower odds in the year prior to have had an inpatient stay (AOR 0.84, CI 0.72-0.99) or a SNF stay (AOR 0.55, CI 0.40-0.75).Based on MCBS data, odds of self-reported inpatient and SNF use were lower among married respondents, while unadjusted odds of outpatient use were higher, compared to currently-not-married beneficiaries.
Vitamin D deficiency has long been associated with reduced immune function that can lead to viral infection. Several studies have shown that Vitamin D deficiency is associated with increases the risk ...of infection with COVID-19. However, it is unknown if treatment with Vitamin D can reduce the associated risk of COVID-19 infection, which is the focus of this study. In the population of US veterans, we show that Vitamin D
and D
fills were associated with reductions in COVID-19 infection of 28% and 20%, respectively (D
Hazard Ratio (HR) = 0.80, 95% CI 0.77, 0.83), D
HR = 0.72, 95% CI 0.65, 0.79. Mortality within 30-days of COVID-19 infection was similarly 33% lower with Vitamin D
and 25% lower with D
(D
HR = 0.67, 95% CI 0.59, 0.75; D
HR = 0.75, 95% CI 0.55, 1.04). We also find that after controlling for vitamin D blood levels, veterans receiving higher dosages of Vitamin D obtained greater benefits from supplementation than veterans receiving lower dosages. Veterans with Vitamin D blood levels between 0 and 19 ng/ml exhibited the largest decrease in COVID-19 infection following supplementation. Black veterans received greater associated COVID-19 risk reductions with supplementation than White veterans. As a safe, widely available, and affordable treatment, Vitamin D may help to reduce the severity of the COVID-19 pandemic.
Since publication of the report by the Panel on Cost-Effectiveness in Health and Medicine in 1996, researchers have advanced the methods of cost-effectiveness analysis, and policy makers have ...experimented with its application. The need to deliver health care efficiently and the importance of using analytic techniques to understand the clinical and economic consequences of strategies to improve health have increased in recent years.
To review the state of the field and provide recommendations to improve the quality of cost-effectiveness analyses. The intended audiences include researchers, government policy makers, public health officials, health care administrators, payers, businesses, clinicians, patients, and consumers.
In 2012, the Second Panel on Cost-Effectiveness in Health and Medicine was formed and included 2 co-chairs, 13 members, and 3 additional members of a leadership group. These members were selected on the basis of their experience in the field to provide broad expertise in the design, conduct, and use of cost-effectiveness analyses. Over the next 3.5 years, the panel developed recommendations by consensus. These recommendations were then reviewed by invited external reviewers and through a public posting process.
The concept of a "reference case" and a set of standard methodological practices that all cost-effectiveness analyses should follow to improve quality and comparability are recommended. All cost-effectiveness analyses should report 2 reference case analyses: one based on a health care sector perspective and another based on a societal perspective. The use of an "impact inventory," which is a structured table that contains consequences (both inside and outside the formal health care sector), intended to clarify the scope and boundaries of the 2 reference case analyses is also recommended. This special communication reviews these recommendations and others concerning the estimation of the consequences of interventions, the valuation of health outcomes, and the reporting of cost-effectiveness analyses.
The Second Panel reviewed the current status of the field of cost-effectiveness analysis and developed a new set of recommendations. Major changes include the recommendation to perform analyses from 2 reference case perspectives and to provide an impact inventory to clarify included consequences.
Background: In the United States, $50,000 per Quality-Adjusted Life-Year (QALY) is a decision rule that is often used to guide interpretation of cost-effectiveness analyses. However, many ...investigators have questioned the scientific basis of this rule, and it has not been updated. Methods: We used 2 separate approaches to investigate whether the $50,000 per QALY rule is consistent with current resource allocation decisions. To infer a lower bound for the decision rule, we estimated the incremental cost-effectiveness of recent (2003) versus pre-"modern era" (1950) medical care in the United States. To infer an upper bound for the decision rule, we estimated the incremental cost-effectiveness of unsubsidized health insurance versus self-pay for nonelderly adults (ages 21-64) without health insurance. We discounted both costs and benefits, following recommendations of the Panel on Cost-Effectiveness in Health and Medicine. Results: Our base case analyses suggest that plausible lower and upper bounds for a cost-effectiveness decision rule are $183,000 per life-year and $264,000 per life-year, respectively. Our sensitivity analyses widen the plausible range (between $95,000 per life-year saved and $264,000 per life-year saved when we considered only health care's impact on quantity of life, and between $109,000 per QALY saved and $297,000 per QALY saved when we considered health care's impact on quality as well as quantity of life) but it remained substantially higher than $50,000 per QALY. Conclusions: It is very unlikely that $50,000 per QALY is consistent with societal preferences in the United States.
Systematic screening and treatment of sleep-disordered breathing (SDB) or obstructive sleep apnea (OSA) in presurgical patients would impose a significant cost burden; therefore, it is important to ...understand whether SDB is associated with worse postoperative outcomes. We sought to determine the impact of SDB on postoperative outcomes in patients undergoing four specific categories of elective surgery (orthopedic, prostate, abdominal, and cardiovascular). The primary outcomes were in-hospital death, total charges, and length of stay (LOS). Two secondary outcomes of interest were respiratory and cardiac complications.
Data were obtained from the Nationwide Inpatient Sample database. Regression models were fitted to assess the independent association between SDB and the outcomes of interest.
The cohort included 1,058,710 hospitalized adult patients undergoing elective surgeries between 2004 and 2008. SDB was independently associated with decreased mortality in the orthopedic (OR, 0.65; 95% CI, 0.45-0.95; P = .03), abdominal (OR, 0.38; 95% CI, 0.22-0.65; P = .001), and cardiovascular surgery groups (OR, 0.54; 95% CI, 0.40-0.73; P < .001) but had no impact on mortality in the prostate surgery group. SDB was independently associated with a small, but statistically significant increase in estimated mean LOS by 0.14 days (P < .001) and estimated mean total charges by $860 (P < .001) in the orthopedic surgery group but was not associated with increased LOS or total charges in the prostate surgery group. In the abdominal and cardiovascular surgery groups, SDB was associated with a significant decrease in adjusted mean LOS of 1.1 days and 0.35 days, respectively (P < .001 for both groups), and adjusted mean total charges of $3,814 and $4,592, respectively (P < .001 for both groups). SDB was independently associated with a significantly increased OR for emergent intubation and mechanical ventilation, noninvasive ventilation, and atrial fibrillation in all four surgical categories. Emergent intubation occurred significantly earlier in the postoperative course in patients with SDB. In the subgroup of patients requiring emergent intubation, LOS, total charges, pneumonias, and in-hospital death were significantly higher in those without SDB.
In this large national study, despite the increased independent association of SDB with postoperative cardiopulmonary complications, the diagnosis of SDB was not independently associated with an increased rate of in-hospital death. SDB had a mixed impact on LOS and total charges by surgical category.