Endocarditis, invasive dental procedures, and antibiotic prophylaxis efficacy in US Medicaid patients.

OBJECTIVE: Antibiotic prophylaxis is recommended before invasive dental procedures to prevent endocarditis in those at high risk, but supporting data are sparse. We therefore investigated any association between invasive dental procedures and endocarditis, and any antibiotic prophylaxis effect on endocarditis incidence. SUBJECTS AND METHODS: Cohort and case-crossover studies were performed on 1,678,190 Medicaid patients with linked medical, dental, and prescription data. RESULTS: The cohort study identified increased endocarditis incidence within 30 days of invasive dental procedures in those at high risk, particularly after extractions (OR 14.17, 95% CI 5.40-52.11, p < 0.0001) or oral surgery (OR 29.98, 95% CI 9.62-119.34, p < 0.0001). Furthermore, antibiotic prophylaxis significantly reduced endocarditis incidence following invasive dental procedures (OR 0.20, 95% CI 0.06-0.53, p < 0.0001). Case-crossover analysis confirmed the association between invasive dental procedures and endocarditis in those at high risk, particularly following extractions (OR 3.74, 95% CI 2.65-5.27, p < 0.005) and oral surgery (OR 10.66, 95% CI 5.18-21.92, p < 0.0001). The number of invasive procedures, extractions, or surgical procedures needing antibiotic prophylaxis to prevent one endocarditis case was 244, 143 and 71, respectively. CONCLUSIONS: Invasive dental procedures (particularly extractions and oral surgery) were significantly associated with endocarditis in high-risk individuals, but AP significantly reduced endocarditis incidence following these procedures, thereby supporting current guideline recommendations.

Antibiotic Prophylaxis Against Infective Endocarditis Before Invasive Dental Procedures.

BACKGROUND: Antibiotic prophylaxis (AP) before invasive dental procedures (IDPs) is recommended to prevent infective endocarditis (IE) in those at high IE risk, but there are sparse data supporting a link between IDPs and IE or AP efficacy in IE prevention. OBJECTIVES: The purpose of this study was to investigate any association between IDPs and IE, and the effectiveness of AP in reducing this. METHODS: We performed a case-crossover analysis and cohort study of the association between IDPs and IE, and AP efficacy, in 7,951,972 U.S. subjects with employer-provided Commercial/Medicare-Supplemental coverage. RESULTS: Time course studies showed that IE was most likely to occur within 4 weeks of an IDP. For those at high IE risk, case-crossover analysis demonstrated a significant temporal association between IE and IDPs in the preceding 4 weeks (OR: 2.00; 95% CI: 1.59-2.52; P = 0.002). This relationship was strongest for dental extractions (OR: 11.08; 95% CI: 7.34-16.74; P < 0.0001) and oral-surgical procedures (OR: 50.77; 95% CI: 20.79-123.98; P < 0.0001). AP was associated with a significant reduction in IE incidence following IDP (OR: 0.49; 95% CI: 0.29-0.85; P = 0.01). The cohort study confirmed the associations between IE and extractions or oral surgical procedures in those at high IE risk and the effect of AP in reducing these associations (extractions: OR: 0.13; 95% CI: 0.03-0.34; P < 0.0001; oral surgical procedures: OR: 0.09; 95% CI: 0.01-0.35; P = 0.002). CONCLUSIONS: We demonstrated a significant temporal association between IDPs (particularly extractions and oral-surgical procedures) and subsequent IE in high-IE-risk individuals, and a significant association between AP use and reduced IE incidence following these procedures. These data support the American Heart Association, and other, recommendations that those at high IE risk should receive AP before IDP.

Electronic phenotyping of health outcomes of interest using a linked claims-electronic health record database: Findings from a machine learning pilot project.

OBJECTIVE: Claims-based algorithms are used in the Food and Drug Administration Sentinel Active Risk Identification and Analysis System to identify occurrences of health outcomes of interest (HOIs) for medical product safety assessment. This project aimed to apply machine learning classification techniques to demonstrate the feasibility of developing a claims-based algorithm to predict an HOI in structured electronic health record (EHR) data. MATERIALS AND METHODS: We used the 2015-2019 IBM MarketScan Explorys Claims-EMR Data Set, linking administrative claims and EHR data at the patient level. We focused on a single HOI, rhabdomyolysis, defined by EHR laboratory test results. Using claims-based predictors, we applied machine learning techniques to predict the HOI: logistic regression, LASSO (least absolute shrinkage and selection operator), random forests, support vector machines, artificial neural nets, and an ensemble method (Super Learner). RESULTS: The study cohort included 32 956 patients and 39 499 encounters. Model performance (positive predictive value [PPV], sensitivity, specificity, area under the receiver-operating characteristic curve) varied considerably across techniques. The area under the receiver-operating characteristic curve exceeded 0.80 in most model variations. DISCUSSION: For the main Food and Drug Administration use case of assessing risk of rhabdomyolysis after drug use, a model with a high PPV is typically preferred. The Super Learner ensemble model without adjustment for class imbalance achieved a PPV of 75.6%, substantially better than a previously used human expert-developed model (PPV = 44.0%). CONCLUSIONS: It is feasible to use machine learning methods to predict an EHR-derived HOI with claims-based predictors. Modeling strategies can be adapted for intended uses, including surveillance, identification of cases for chart review, and outcomes research.

Two-Year Costs and Quality in the Comprehensive Primary Care Initiative.

BACKGROUND: The 4-year, multipayer Comprehensive Primary Care Initiative was started in October 2012 to determine whether several forms of support would produce changes in care delivery that would improve the quality and reduce the costs of care at 497 primary care practices in seven regions across the United States. Support included the provision of care-management fees, the opportunity to earn shared savings, and the provision of data feedback and learning support. METHODS: We tracked changes in the delivery of care by practices participating in the initiative and used difference-in-differences regressions to compare changes over the first 2 years of the initiative in Medicare expenditures, health care utilization, claims-based measures of quality, and patient experience for Medicare fee-for-service beneficiaries attributed to initiative practices and a group of matched comparison practices. RESULTS: During the first 2 years, initiative practices received a median of $115,000 per clinician in care-management fees. The practices reported improvements in approaches to the delivery of primary care in areas such as management of the care of high-risk patients and enhanced access to care. Changes in average monthly Medicare expenditures per beneficiary did not differ significantly between initiative and comparison practices when care-management fees were not taken into account (-$11; 95% confidence interval [CI], -$23 to $1; P=0.07; negative values indicate less growth in spending at initiative practices) or when these fees were taken into account ($7; 95% CI, -$5 to $19; P=0.27). The only significant differences in other measures were a 3% reduction in primary care visits for initiative practices relative to comparison practices (P<0.001) and changes in two of the six domains of patient experience--discussion of decisions regarding medication with patients and the provision of support for patients taking care of their own health--both of which showed a small improvement in initiative practices relative to comparison practices (P=0.006 and P<0.001, respectively). CONCLUSIONS: Midway through this 4-year intervention, practices participating in the initiative have reported progress in transforming the delivery of primary care. However, at this point these practices have not yet shown savings in expenditures for Medicare Parts A and B after accounting for care-management fees, nor have they shown an appreciable improvement in the quality of care or patient experience. (Funded by the Department of Health and Human Services, Centers for Medicare and Medicaid Services; ClinicalTrials.gov number, NCT02320591.).

Hospital Characteristics and the Agency for Healthcare Research and Quality Inpatient Quality Indicators: A Systematic Review.

BACKGROUND: The Agency for Healthcare Research and Quality Inpatient Quality Indicators (IQIs) include inpatient mortality for selected procedures and medical conditions. They have assumed an increasingly prominent role in hospital comparisons. Healthcare delivery and policy-related decisions need to be driven by reliable research that shows associations between hospital characteristics and quality of inpatient care delivered. OBJECTIVES: To systematically review the literature on associations between hospital characteristics and IQIs. METHODS: We systematically searched PubMed and gray literature (2000-2012) for studies relevant to 14 hospital characteristics and 17 IQIs. We extracted data for study characteristics, IQIs analyzed, and hospital characteristics (e.g., teaching status, bed size, patient volume, rural vs. urban location, and nurse staffing). RESULTS: We included 16 studies, which showed few significant associations. Four hospital characteristics (higher hospital volume, higher nurse staffing, urban vs. rural status, and higher hospital financial resources) had statistically significant associations with lower mortality and selected IQIs in approximately half of the studies. For example, there were no associations between nurse staffing and four IQIs; however, approximately 50% of studies showed a statistically significant relationship between nurse staffing and lower mortality for six IQIs. For two hospital characteristics-higher bed size and disproportionate share percentage-all statistically significant associations had higher mortality. Five hospital characteristics (teaching status, system affiliation, ownership, minority-serving hospitals, and electronic health record status) had some studies with significantly positive and some with significantly negative associations, and many studies with no association. CONCLUSIONS: We found few associations between hospital characteristics and mortality IQIs. Differences in study methodology, coding across hospitals, and hospital case-mix adjustment may partly explain these results. Ongoing research will evaluate potential mechanisms for the identified associations.

Variable-ratio matching with fine balance in a study of the Peer Health Exchange.

In some observational studies of treatment effects, matched samples are created so treated and control groups are similar in terms of observable covariates. Traditionally, such matched samples consist of matched pairs. However, alternative forms of matching may have desirable features. One strategy that may improve efficiency is to match a variable number of control units to each treated unit. Another strategy to improve balance is to adopt a fine balance constraint. Under a fine balance constraint, a nominal covariate is exactly balanced, but it does not require individually matched treated and control subjects for this variable. Here, we propose a method to allow for fine balance constraints when each treated unit is matched to a variable number of control units, which is not currently possible using existing matching network flow algorithms. Our approach uses the entire number to first determine the optimal number of controls for each treated unit. For each stratum of matched treated units, we can then apply a fine balance constraint. We then demonstrate that a matched sample for the evaluation of the Peer Health Exchange, an intervention in schools designed to decrease risky health behaviors among youths, using a variable number of controls and fine balance constraint is superior to simply using a variable-ratio match. Copyright © 2015 John Wiley & Sons, Ltd.

The effects of mental health parity on spending and utilization for bipolar, major depression, and adjustment disorders.

OBJECTIVE: The Mental Health Parity and Addiction Equity Act requires insurance parity for mental health/substance use disorder and general medical services. Previous research found that parity did not increase mental health/substance use disorder spending and lowered out-of-pocket spending. Whether parity's effects differ by diagnosis is unknown. The authors examined this question in the context of parity implementation in the Federal Employees Health Benefits (FEHB) Program. METHOD: The authors compared mental health/substance use disorder treatment use and spending before and after parity (2000 and 2002, respectively) for two groups: FEHB enrollees diagnosed in 1999 with bipolar disorder, major depression, or adjustment disorder (N=19,094) and privately insured enrollees unaffected by the policy in a comparison national sample (N=10,521). Separate models were fitted for each diagnostic group. A difference-in-difference design was used to control for secular time trends and to better reflect the specific impact of parity on spending and utilization. RESULTS: Total spending was unchanged among enrollees with bipolar disorder and major depression but decreased for those with adjustment disorder (-$62, 99.2% CI=-$133, -$11). Out-of-pocket spending decreased for all three groups (bipolar disorder: -$148, 99.2% CI=-$217, -$85; major depression: -$100, 99.2% CI=-$123, -$77; adjustment disorder: -$68, 99.2% CI=-$84, -$54). Total annual utilization (e.g., medication management visits, psychotropic prescriptions, and mental health/substance use disorder hospitalization bed days) remained unchanged across all diagnoses. Annual psychotherapy visits decreased significantly only for individuals with adjustment disorders (-12%, 99.2% CI=-19%, -4%). CONCLUSIONS: Parity implemented under managed care improved financial protection and differentially affected spending and psychotherapy utilization across groups. There was some evidence that resources were preferentially preserved for diagnoses that are typically more severe or chronic and reduced for diagnoses expected to be less so.

Using Multiple Control Groups and Matching to Address Unobserved Biases in Comparative Effectiveness Research: An Observational Study of the Effectiveness of Mental Health Parity.

Studies of large policy interventions typically do not involve randomization. Adjustments, such as matching, can remove the bias due to observed covariates, but residual confounding remains a concern. In this paper we introduce two analytical strategies to bolster inferences of the effectiveness of policy interventions based on observational data. First, we identify how study groups may differ and then select a second comparison group on this source of difference. Second, we match subjects using a strategy that finely balances the distributions of key categorical covariates and stochastically balances on other covariates. An observational study of the effect of parity on the severely ill subjects enrolled in the Federal Employees Health Benefits (FEHB) Program illustrates our methods.

Alternative methods for testing treatment effects on the basis of multiple outcomes: simulation and case study.

In clinical trials multiple outcomes are often used to assess treatment interventions. This paper presents an evaluation of likelihood-based methods for jointly testing treatment effects in clinical trials with multiple continuous outcomes. Specifically, we compare the power of joint tests of treatment effects obtained from joint models for the multiple outcomes with univariate tests based on modeling the outcomes separately. We also consider the power and bias of tests when data are missing, a common feature of many trials, especially in psychiatry. Our results suggest that joint tests capitalize on the correlation of multiple outcomes and are more powerful than standard univariate methods, especially when outcomes are missing completely at random. When outcomes are missing at random, test procedures based on correctly specified joint models are unbiased, while standard univariate procedures are not. Results of a simulation study are reported, and the methods are illustrated in an example from the Clinical Antipsychotic Trials of Intervention Effectiveness for schizophrenia.

An investigation into differential diagnosis of pulp and periapical pain: a PennEndo database study.

Nine hundred fifty-one emergency and 997 nonemergency patients seeking endodontic treatment were the basis of this study. Variables of interest were 10 pain descriptors, percussion and palpation tests, causative factors, and paired pulpal and periapical diagnoses. A higher number of patients suffering from symptomatic pulpal conditions sought emergency care. Odds of caries being a causative factor were high in symptomatic pulps compared with asymptomatic pulpal and periapical conditions. Higher odds ratios were obtained for sharp pain in symptomatic pulps versus symptomatic periapical conditions. Conversely, odds ratios for dull pain were higher in symptomatic periapical conditions compared with asymptomatic periapical conditions. Percussion and palpation tests were significant in differentially diagnosing between pulpal and periapical conditions. In conclusion, caries was associated with painful pulpitis. The results confirm the differential diagnostic power of sharp and dull pain and percussion and palpation tests. Several symptoms previously believed to have differential diagnostic power were found insignificant.