Data Science for Child Health
Overview of the field of data science as it might be applied to pediatric domains. Bennett 2019 (J. Pediatrics) | PubMed PMC6486872 | Author Search
Data Issues in Child Health
Special methods of handling data used in the health care of infants, children, and adolescents
Distinct Newborn Identification Requirement
From the Joint Commission, about a National Patient Safety Goal (NPSG.01.01.01) that goes into effect 1/1/19, on more distinguishable newborn naming conventions, for example, using the mother’s first and last names and the newborn’s gender (plus added letters for multiple births) to create a temporary name. (R3 Report. June 25, 2018;7:1-2).
Pediatric Weight Errors and Resultant Medication Dosing Errors in the Emergency Department
A retrospective study of 79,000 ED encounters at a children’s hospital and two general hospitals. The intent of the study was to characterize the frequency of weight errors and to determine of the children’s hospital was any better at correcting errors than the general hospitals. The findings were that weight errors were uncommon (0.63% of all weights, as defined by the weight being a new extreme value on the growth chart) in the 3 EDs, but they led to identifiable weight-based medication-dosing errors with the potential to cause harm. The rates of error where similar across hospitals, and it looked like the children’s hospital was slightly better at intercepting errors once they were committed. Common weight errors included the weight in pounds being substituted for the weight in kilograms and decimal placement errors.
Tenfold Medication Errors: 5 Years’ Experience at a University-Affiliated Pediatric Hospital
A study of pediatric inpatient safety reports. From the abstract: “From 6643 medication-related safety reports, 252 10-fold medication errors were identified at a mean reporting rate of 0.062 per 100 total patient days. Morphine was the most frequently reported medication, and opioids were the most frequently reported drug class. Twenty-two reports described patient harm. Intravenous formulations, paper ordering, and drug-delivery pumps were frequent error enablers. Errors of dose calculation, documentation of decimal points, and confusion with zeroes were frequent contributing causes to 10-fold medication error.”
[su_cite_pediatrics url_fragment = ‘129/5/916’ author = ‘Doherty’ year = ‘2012’] | PubMed 22473367 | Author Search
The Effect of an Electronic Health Record–based Tool on Abnormal Pediatric Blood Pressure Recognition
Evaluation of whether provider recognition of abnormal BP (greater than 90th percentile) differed before versus after the introduction of an app that extracts age, sex, height and BP data from the EHR to calculate and track a patient’s BP percentile longitudinally. The app was based on the Substitutable Medical Applications & Reusable Technology (SMART) platform and is available In the SMARTApp Gallery. Examining ~79,000 records of outpatients (primary care, endocrinology, cardiology, nephrology clinics), of which ~3500 had elevated blood pressure, showed that abnormal BP was recognized in 4.9% of visits before the app was available and 7.1% of visits afterwards. The app was used in 13% of encounters where an elevated BP was present; significantly, when the app was used, recognition of elevated BP was much higher (OR 3.17, CI 2.29-4.41).
Automated Identification of Implausible Values in Growth Data from Pediatric Electronic Health Records
Reports the development of an automated method for identifying implausible values in pediatric EHR growth (weight and height) data, tested via data points collected in the primary care environment on over 280,000 patiets. The method compares each measurement’s z-score to a weighted moving average of prior measurements. The method had a sensitivity of 97% and a specificity of 90% for identifying implausible values compared to physician judgment, and identified almost all simulated errors.
Pediatric Complex Chronic Conditions Classification System Version 2: Updated for ICD-10 and Complex Medical Technology Dependence and Transplantation
An ICD-10 update of the pediatric complex chronic conditions (CCC) classification system from 2000. The system includes diagnostic and procedural codes that incorporate a new neonatal CCC category as well as domains of complexity arising from technology dependence or organ transplantation. Linked electronic supplementary material provides SAS and Stata code, plus tabular information on codes.
Feudtner 2014 (BMC Pediatrics) | PubMed 25102958 | Author Search
Real-time Electronic Medical Record Alerts Increase High blood Pressure Recognition in Children
Small study of high-blood-pressure alert effectiveness suggesting that such an alert raised recognition of hypertension from 12% to 42%.
Brady 2015 (The Journal of Emergency Medicine) | PubMed 25416626 | Author Search
Recognition of Elevated Blood Pressure in an Outpatient Pediatric Tertiary Care Setting
Another study suggesting we are not very good at officially diagnosing elevated blood pressure in children when the record shows BP readings that are repeatedly elevated. Of 29,000 records retrospectively reviewed, only about 1 in 6 of those with repeatedly high blood pressure values were diagnosed in the encounter diagnoses or problem list.
Beacher 2015 (J. Pediatrics) | PubMed 25919733 | Author Search
The Transition to ICD-10-CM: Challenges for Pediatric Practice
An evaluation of diagnostic codes from Medicaid claims data that suggests that the use of the CMS General Equivalency Mappings will result in loss or obfuscation of clinical concepts. 40% of ICD-9 diagnostic codes in this study were the undefined “999.99,” which suggests that conclusions from analysis of this data set may be limited due to the poor quality of the coding efforts upstream of these data. Given that ICD-9 itself obfuscates much clinical detail in the first place, the jury is still out on whether ICD-10 is better or worse for pediatric care than its predecessor.
[su_cite_pediatrics url_fragment = ‘134/1/31’ author = ‘Caskey’ year = ‘2014’] | PubMed 24918217 | Author Search