Search

We sought to determine whether machine learning and natural language processing (NLP) applied to electronic medical records could improve performance of automated health-care claims-based algorithms to identify anaphylaxis events using data on 516 patients with outpatient, emergency department, or inpatient anaphylaxis diagnosis codes during 2015–2019 in 2 integrated health-care institutions in the Northwest United States. We used one site’s manually reviewed gold-standard outcomes data for model development and the other’s for external validation based on cross-validated area under the receiver operating characteristic curve (AUC), positive predictive value (PPV), and sensitivity. In the development site 154 (64%) of 239 potential events met adjudication criteria for anaphylaxis compared with 180 (65%) of 277 in the validation site. Logistic regression models using only structured claims data achieved a cross-validated AUC of 0.58 (95% CI: 0.54, 0.63). Machine learning improved cross-validated AUC to 0.62 (0.58, 0.66); incorporating NLP-derived covariates further increased cross-validated AUCs to 0.70 (0.66, 0.75) in development and 0.67 (0.63, 0.71) in external validation data. A classification threshold with cross-validated PPV of 79% and cross-validated sensitivity of 66% in development data had cross-validated PPV of 78% and cross-validated sensitivity of 56% in external data. Machine learning and NLP-derived data improved identification of validated anaphylaxis events.

Objective To evaluate the association between obstructive sleep apnea (OSA) and risk of motor vehicle accident (MVA). Methods We conducted a cohort study at Kaiser Permanente Washington using electronic health plan data and linked Washington State Department of Transportation MVA records. We included persons 18–79 years of age during 2005–2014. OSA was ascertained via diagnosis codes. The primary outcome, first MVA during cohort follow-up, was ascertained from state MVA records. Risk factors for MVAs, including medical conditions and medication use, were ascertained from health plan data. Multivariable Cox proportional hazards models were used to estimate the adjusted hazard ratio (HR) and 95% confidence interval (CI) for the association between OSA and study outcomes. Results Among the 879,547 eligible persons, the unadjusted rate of MVA in those with and without OSA was 238 and 229 per 10,000 person-years, respectively. A diagnosis of OSA was associated with a 17% increased risk of MVA (adjusted HR = 1.17; 95% CI: 1.13 to 1.20). Conclusion In this large population-based study, a diagnosis of OSA was associated with a modestly increased risk of MVA.

Background A distributed data network approach combined with distributed regression analysis (DRA) can reduce the risk of disclosing sensitive individual and institutional information in multicenter studies. However, software that facilitates large-scale and efficient implementation of DRA is limited. Objective This study aimed to assess the precision and operational performance of a DRA application comprising a SAS-based DRA package and a file transfer workflow developed within the open-source distributed networking software PopMedNet in a horizontally partitioned distributed data network. Methods We executed the SAS-based DRA package to perform distributed linear, logistic, and Cox proportional hazards regression analysis on a real-world test case with 3 data partners. We used PopMedNet to iteratively and automatically transfer highly summarized information between the data partners and the analysis center. We compared the DRA results with the results from standard SAS procedures executed on the pooled individual-level dataset to evaluate the precision of the SAS-based DRA package. We computed the execution time of each step in the workflow to evaluate the operational performance of the PopMedNet-driven file transfer workflow. Results All DRA results were precise ( Conclusions We integrated a SAS-based DRA package with PopMedNet and successfully tested the new capability within an active distributed data network. The study demonstrated the validity and feasibility of using DRA to enable more privacy-protecting analysis in multicenter studies.

This study examines medication-related self-efficacy in a linguistically diverse group of patients with diabetes, hypertension, and elevated cholesterol. A telephone survey of 509 adults conducted in six languages (English, Spanish, Korean, Vietnamese, Mandarin, and Cantonese) was analyzed. Self-efficacy was assessed with the overall Medication Understanding and Use Self-Efficacy (MUSE) score and its two subscale scores on taking medication and learning about medications. Compared with English proficient (EP) patients, patients with limited English proficiency (LEP) had a lower mean learning self-efficacy subscale score (LEP: 14.5, EP: 15.4; plt;.001) and no difference in the mean taking self-efficacy subscale score (LEP: 14.4, EP: 14.6; p=.40). Receiving verbal medication information (VMI) from providers modified the relationship between LEP status and learning self-efficacy. In conclusion, among patients with chronic illnesses, LEP patients had lower medication-related self-efficacy scores than EP patients, which may put them at greater risk for medication taking errors and lower adherence.

Summary Objective: We evaluated the role of home monitoring, communication with pharmacists, medication intensification, medication adherence and lifestyle factors in contributing to the effectiveness of an intervention to improve blood pressure control in patients with uncontrolled essential hypertension. Methods: We performed a mediation analysis of a published randomized trial based on the Chronic Care Model delivered over a secure patient website from June 2005 to December 2007. Study arms analyzed included usual care with a home blood pressure monitor and usual care with home blood pressure monitor and web-based pharmacist care. Mediator measures included secure messaging and telephone encounters; home blood pressure monitoring; medications intensification and adherence and lifestyle factors. Overall fidelity to the Chronic Care Model was assessed with the Patient Assessment of Chronic Care (PACIC) instrument. The primary outcome was percent of participants with blood pressure (BP) Results: At 12 months follow-up, patients in the web-based pharmacist care group were more likely to have BP Conclusions: The effect of web-based pharmacist care on improved blood pressure control was explained in part through a combination of home blood pressure monitoring, secure messaging, and antihypertensive medication intensification. Citation: Ralston JD, Cook AJ, Anderson ML, Catz SL, Fishman PA, Carlson J, Johnson R, Green BB. Home blood pressure monitoring, secure electronic messaging and medication intensification for improving hypertension control: a mediation analysis. Appl Clin Inf 2014; 5: 232–248 http://dx.doi.org/10.4338/ACI-2013-10-RA-0079

Background The US Preventive Services Taskforce (USPSTF) recommends routine lipid screening beginning age 35 for men [1]. For women age 20 and older, as well as men age 20–34, screening is recommended if cardiovascular risk factors are present. Prior research has focused on underutilization but not overuse of lipid testing. The objective is to document over- and under-use of lipid testing in an insured population of persons at low, moderate and high cardiovascular disease (CVD) risk for persons not already on statins. Methods The study is a retrospective cohort study that included all adults without prior CVD who were continuously enrolled in a large integrated healthcare system from 2005 to 2010. Measures included lipid test frequency extracted from administrative data and Framingham cardiovascular risk equations applied using electronic medical record data. Five year lipid testing patterns were examined by age, sex and CVD risk. Generalized linear models were used to estimate the relative risk for over testing associated with patient characteristics. Results Among males and females for whom testing is not recommended, 35.8 % and 61.5 % received at least one lipid test in the prior 5 years and 8.4 % and 24.4 % had two or more. Over-testing was associated with age, race, comorbidity, primary care use and neighborhood income. Among individuals at moderate and high-risk (not already treated with statins) and for whom screening is recommended, between 21.4 % and 25.1 % of individuals received no screening in the prior 5 years. Conclusions Based on USPSTF lipid screening recommendations, this study documents substantial over-testing among individuals with low CVD risk and under-testing among individuals with moderate to high-risk not already on statins. Opportunity exists to better focus lipid screening efforts appropriate to CVD risk.

Health systems are increasingly implementing remote telephone and Internet refill systems to enhance patient access to medication refills. Remote refill systems may provide an effective approach for improving medication non-adherence, but more research is needed among patients with limited English proficiency with poor access to remote refill systems.To compare the use of remote medication refill systems among limited-English-proficiency (LEP) and English-proficient (EP) patients with chronic conditions.Cross-sectional survey in six languages/dialects (English, Cantonese, Mandarin, Korean, Vietnamese, and Spanish) of 509 adults with diabetes, hypertension, or hyperlipidemia. Primary study outcomes were self-reported use of 1) Internet refills, 2) telephone refills, and 3) any remote refill system. LEP was measured by patient self-identification of a primary language other than English and a claims record of use of an interpreter. Other measures were age, gender, education, years in the U.S., insurance, health status, chronic conditions, and number of prescribed medications. Analyses included multivariable logistic regression weighted for survey non-response.Overall, 33.1 % of patients refilled their medications by telephone and 31.6 % by Internet. Among LEP patients (n = 328), 31.5 % refilled by telephone and 21.2 % by Internet, compared with 36.7 % by telephone and 52.7 % by Internet among EP patients (n = 181). Internet refill by language groups were as follows: English (52.7 %), Cantonese (34.9 %), Mandarin (17.4 %), Korean (16.7 %), Vietnamese (24.4 %), and Spanish (12.6 %). Compared to EP patients, LEP patients had lower use of any remote refill system (adjusted odds ratio [AOR] 0.18; p 0.001), CONCLUSIONS: LEP patients are significantly less likely than EP patients to use any remote medication refill system. Increased reliance on current systems for remote medication refills may increase disparities in health outcomes affecting LEP patients with poor access to telephone and Internet medication refills.

Background/Aims In the Electronic Communications and Home Blood Pressure Trial (e-BP) patients with uncontrolled blood pressure (BP) were registered to use an existing patient shared electronic health record (EHR) and secure e-mail and randomly assigned to: (1) usual care (UC); (2) home BP monitoring (BPM) and website training; or (3) this plus pharmacist team-care delivered via the web (Pharm). At the end of intervention and one year later (1 and 2 years after randomization) Pharm patients were more likely to have controlled BP. The objective here was to determine if BP control improvements persisted longer term.

Background: Abuse of over-the-counter (OTC) products, such as diet pills and laxatives, for weight control by adolescents is well-documented and can precipitate serious medical conditions. Yet only a small percentage of youth with disordered weight control behaviors receive treatment. The objective of this study was to examine how often clinicians communicate with youth with symptoms consistent with abuse of OTC products for weight control about possible use of these products. We used electronic medical records and administrative claims for services for 53,229 12 to 17 year old patients receiving care from an integrated health system in the U.S. Northwest from August 2007 to December 2010. We examined electronic text of clinical notes to identify encounters in which the clinician noted one of 10 metabolic conditions potentially associated with abuse of OTC products (diet pills, laxatives, diuretics, ipecac, orlistat, and alli®) for weight control and then assessed whether clinicians noted communication with adolescent patients about possible use of OTC products for weight control. Results: We identified 130 (0.2% of sample) patients with clinical notes indicating one or more of the metabolic conditions. In clinical notes for only four (3.1%) of these patients did clinicians document suspicion or communication about possible abuse of the OTC products. All four had a previous eating disorder diagnosis. In the 12 months subsequent to the clinical encounter in which a metabolic disturbance was identified, medical notes for only three (2.3%) of the 130 patients indicated clinician suspicion or communication about possible abuse of these products or an eating disorder. Conclusions: Clinicians are missing a critical window of opportunity to query adolescents when presenting with suspicious metabolic disturbances about possible abuse of OTC products for weight control. Clinicians may need more training to detect OTC product abuse, and electronic medical records should prompt more thorough enquiry.

Formaldehyde decay was followed by IR emission at 3.5 μm behind reflected shocks at total concentrations near 2.4 and 4.5×1018 molecule/cm3 in a 7.6 cm diameter shock tube. Data were obtained at 0.5 μs intervals over a temperature range of 1700–2710 K for mixtures containing 0.1%, 0.5%, and 1.0% CH2O in Ar. Exponential decay was observed in all cases; at lower temperatures there was evidence of an induction time prior to the onset of the decay. An apparent rate constant was defined as ka=−[d(In signal)/dt]/[M]; ka was observed to be independent of the total concentration but to decrease with decreasing CH2O concentration. These data were analyzed in terms of the mechanism: C H 2 O + M = HCO + H + M (1) C H 2 O + H = HCO + H 2 (2) HCO + M = H + CO + M (3) HCO + H = H 2 + CO ( 4 ) (4) Using literature values for k2, k3, and k4, a reasonable fit to all the data, both in terms of temperature and concentration dependence, was obtained with k1=6.0×10−7 exp(−364 kJ/RT) cm3 molecule−1 s−1. Error considerations suggest this value of k1 is accurate to within a factor of three between 1800–2500 K. However, the uncertainty in the activation energy may be as high as 40–50 kJ/mole. Evidence is also presented to suggest that the reaction C H 2 O + M = CO + H 2 + M ( 1 ′ ) does not play a significant role in the high temperature pyrolysis of formaldehyde.

Reflected shock waves were used to heat N 2 O/C 2 H 6 /CO/Ar and N 2 O/CH 4 /CO/Ar mixtures to temperatures of 1875–2855 K at total concentrations of 5 × 10 18 cm −3 . Oxygen atom production was monitored via the flame band emission at 450 nm, and CO 2 production was observed at 4.27 μm. These data were then compared to the results of numerical integration of the rate equations, using rate constant and mechanistic information obtained in earlier studies of H 2 and CH 2 O oxidation under similar conditions. For the C 2 H 6 mixture, it was possible to achieve good agreement between these calculations and the observed data using only one additional reaction O+CH 3 →CH 2 O+OH The calculations were insensitive to the precise value of the rate constant used, but the results are consistent with recent high temperature literature values. For the CH 4 mixture, the best agreement was achieved using a recent high temperature value for O+CH 4 →CH 3 +OH The data indicated that this reaction has a markedly non-Arrhenius rate constant; use of a value based upon low temperature data gives results inconsistent with those observed. The CH 4 system was insensitive to any other methane reaction rate constant. There was some experimental evidence to suggest an unsuspected complexity in methyl radical decay channels at the lowest temperatures observed. The mechanism/rate constant combination used here was then applied to the analysis of literature data for CH 3 oxidation by O 2 . This analysis was designed to obtain a value for the rate constant of CH 3 +O 2 →CH 2 O+OH Results indicated that the most recent high temperature studies yield values consistent with the present analysis.

Reflected shock waves were used to heat a variety of formaldehyde mixtures to temperatures of 1600–3000 K at total concentrations near 5 × 1018/cm3. Formaldehyde decay was monitored by infrared (IR) emission at 3.50 μm using both O2 and N2O as the oxidant in mixtures highly diluted with argon. Other mixtures with added CO were monitored for oxygen-atom production via the flame-band emission at 450 nm and for CO2 production via IR emission at 4.27 μm. These data were then compared to the results obtained from numerical integration of the rate equations based on a likely mechanism. It was possible to achieve good agreement with the formaldehyde reaction-rate constants shown as follows: CH 2 + M = HCO + H + M k 1 =5.5×10 −8 exp(−339 kJ/RT ) cm 3 sec −1 The values listed for k3 and k4 are taken directly from the low-temperature studies of these reactions; the k1 and k2 values also fit recent experiments in this laboratory on CH2O pyrolysis. The N2O experiments also served as a check on the rate-constant assignment for the reaction: H + N 2 O = N 2 + OH k 19 =1.5×10 −9 exp(−92 kJ/RT )cm 3 sec − Reasonable agreement was achieved here. Sensitivity studies suggest these rate constants are probably accurate to within a factor of 2 at 1700–2500 K. The values of k1 and k3 obtained in this work are at least an order of magnitude lower than the values frequently used in current modeling studies of methane oxidation. The implications of these differences are discussed in light of recent suggestions that C2 species participate in methane oxidation.

This study was conducted to provide current information on the effectiveness of water treatment chemicals and filters for control of Giardia cysts in areas where treated water is not available. Four filters and seven chemical treatments were evaluated for both clear and turbid water at 10 degrees C. Three contact disinfection devices were also tested for cyst inactivation. Filters were tested with 1-liter volumes of water seeded with 3 x 10(4) cysts of G. lamblia produced in gerbils inoculated with in vitro cultured trophozoites; the entire volume of filtrate was examined for cyst passage. Chemical treatments were evaluated at concentrations specified by the manufacturer and for contact times that might be expected of hikers (30 minutes) and campers (eight hours, i.e., overnight). Two of the four filter devices tested were 100 percent effective for Giardia cyst removal. Of the other two filters, one was 90 percent effective and the other considerably less effective. Among the seven disinfection treatments, the iodine-based chemicals were all significantly more effective than the chlorine-based chemicals. None of the chemical treatments achieved 99.9 percent cyst inactivation with only 30-minute contact. After an eight-hour contact each of the iodine but none of the chlorine preparations achieved at least 99.9 percent cyst inactivation. None of the contact disinfection devices provided appreciable cyst inactivation. Heating water to at least 70 degrees C for 10 minutes was an acceptable alternative treatment.