Assessment of International Classification of Diseases, Tenth Revision, Clinical Modification (ICD-10-CM) Code Assignment Validity for Case Finding of Medication-related Hypoglycemia Acute Care Visits Among Medicare Beneficiaries

OBJECTIVE

Administrative claims are commonly relied upon to identify hypoglycemia. We assessed validity of 14 International Classification of Diseases, Tenth Revision, Clinical Modification diagnosis code assignments to identify medication-related hypoglycemia leading to acute care encounters.

RESEARCH DESIGN AND METHODS

A multisite, retrospective medical record review study was conducted in a sample of Medicare beneficiaries prescribed outpatient diabetes medications and who received hospital care between January 1, 2016 and September 30, 2017. Diagnosis codes were validated with structured medical record review using prespecified criteria (clinical presentation, blood glucose values, and treatments for hypoglycemia). Sensitivity, specificity, and positive and negative predictive value (PPV, NPV) were calculated and adjusted using sampling weights to correct for partial verification bias.

RESULTS

Among 990 encounters (496 cases, 494 controls), hypoglycemia codes demonstrated moderate PPV (69.2%; 95% confidence interval: 65.0-73.0) and moderate sensitivity (83.9%; 95% confidence interval: 70.0-95.5). Codes performed better at identifying hypoglycemic events among emergency department/observation encounters compared with hospitalizations (PPV 92.9%, sensitivity 100.0% vs. PPV 53.7%, sensitivity 71.0%). Accuracy varied by diagnosis position, especially for hospitalizations, with PPV of 95.6% versus 46.5% with hypoglycemia in primary versus secondary positions. Use of adverse event/poisoning codes did not improve accuracy; reliance on these codes alone would have missed 97% of true hypoglycemic events.

CONCLUSIONS

Accuracy of International Classification of Diseases, Tenth Revision codes in administrative claims to identify medication-related hypoglycemia varied substantially by encounter type and diagnosis position. Consideration should be given to the trade-off between PPV and sensitivity when selecting codes, encounter types, and diagnosis positions to identify hypoglycemia.

Excess death estimates in patients with end-stage renal disease - United States, February-August 2020

This article describes an estimated 6953–10 316 excess deaths among individuals with end-stage renal disease during the early months of the COVID-19 pandemic in February 2020–August 2020 (compared to death rates prior to the pandemic). Notably, the estimated number of excess deaths was reported as 10.8–16.6 per 1000 individuals on dialysis and 2.6–5.5 per 1000 individuals with a prior kidney transplantation. Although not adjusted for confounders, these data suggest that the immunosuppression associated with kidney transplantation is not a dominant determinant of outcomes associated with COVID-19 in this population.

Abortion Facility Closings and Abortion Rates in Texas

From 2004 to 2014, the overall abortion rate in Texas fell by almost a third from 10.7 to 7.2 abortions per 1000 women aged 10 to 49 years. During this same period, the number of abortion clinics operating at least 6 months in the year fell from 40 to 27. We examined the relationship between the abortion rate and the proximity of abortion facilities. We matched annual, county-level data on abortion rates in Texas from 2004 through 2014 with the distance from the county centroids to the nearest abortion facility in operation. Linear regressions were used to estimate the association between abortion rates and proximity to abortion facilities. The regressions controlled for county-level and state-level characteristics as well as the availability of abortion services in neighboring US states and Mexico. We found that a 100-mile increase in distance to the nearest abortion facility was associated with a 10% decrease in the overall abortion rate. The relationship appeared to be driven largely by distances of 200 miles or more. The overall relationship was generally present for whites and blacks, whereas the pattern was less clear for Hispanics. The analysis indicated that the overall association was driven largely by women aged 20 to 34 years. Decreased access to abortion facilities was associated with decreases in the abortion rate, yet the relationship varied by race/ethnicity and age. As such, regulations that affect the operational status of abortion facilities likely have differential effects on women.

The Impact of the Teen Outreach Program on Sexual Intentions and Behaviors

PURPOSE

We evaluated the impact of a positive youth development program on adolescent pregnancy, sexual behavior, risky sex, and intentions in nonmetropolitan Florida high schools.

METHODS

Between 2012 and 2014, the Teen Outreach Program (TOP) was compared to standard school curriculum content using a cluster randomized controlled trial design with 7,976 youth in two cohorts. The majority of youth were 14 years old and in the ninth grade at baseline. Treatment group youth received TOP in health-related classes. After using multiple imputation to account for missing data, we analyzed baseline and follow-up survey data using generalized linear mixed-effects models with logit link function.

RESULTS

In the cohort 1 sample, compared to the control condition, males and females receiving TOP showed lower odds of engaging in recent sex (odds ratio [OR], .71; 95% confidence interval [CI]: .58-.86) compared to control males and females. Cohort 1 treatment females who did engage in recent sex were less likely to have risky sex (OR, .54; 95% CI: .32-.89). There were fewer significant findings in cohort 2, though TOP females and combined gender had lower odds of risky sex intentions (OR, .53; 95% CI: .33-.84 and OR, .65; 95% CI: .44-.96, respectively). Overall, cohort 1 females in the TOP condition were the group most likely to benefit from TOP.

CONCLUSIONS

Consistent with previous research, TOP was more effective regarding sexual health outcomes among female versus male youth; this was especially true for the outcome of risky sex. However, results were not consistent across cohorts, prompting questions for future research.

Live vaccine strain of Francisella tularensis: infection and immunity in mice

The live vaccine strain (LVS) of Francisella tularensis caused lethal disease in several mouse strains. Lethality depended upon the dose and route of inoculation. The lethal dose for 50% of the mice (LD50) in four of six mouse strains (A/J, BALB/cHSD, C3H/HeNHSD, and SWR/J) given an intraperitoneal (i.p.) inoculation was less than 10 CFU. For the other two strains tested, C3H/HeJ and C57BL/6J, the i.p. log LD50 was 1.5 and 2.7, respectively. Similar susceptibility was observed in mice inoculated by intravenous (i.v.) and intranasal (i.n.) routes: in all cases the LD50 was less than 1,000 CFU. Regardless of the inoculation route (i.p., i.v., or i.n.), bacteria were isolated from spleen, liver, and lungs within 3 days of introduction of bacteria; numbers of bacteria increased in these infected organs over 5 days. In contrast to the other routes of inoculation, mice injected with LVS intradermally (i.d.) survived infection: the LD50 of LVS by this route was much greater than 10(5) CFU. This difference in susceptibility was not due solely to local effects at the dermal site of inoculation, since bacteria were isolated from the spleen, liver, and lungs within 3 days by this route as well. The i.d.-infected mice were immune to an otherwise lethal i.p. challenge with as many as 10(4) CFU, and immunity could be transferred with either serum, whole spleen cells, or nonadherent spleen cells (but not Ig+ cells). A variety of infectious agents induce different disease syndromes depending on the route of entry. Francisella LVS infection in mice provides a model system for analysis of locally induced protective effector mechanisms.