Racial and ethnic disparity in clinical outcomes among patients with confirmed COVID-19 infection in a large US electronic health record database

BACKGROUND

Racial and ethnic minority groups have been disproportionately affected by the US coronavirus disease 2019 (COVID-19) pandemic; however, nationwide data on COVID-19 outcomes stratified by race/ethnicity and adjusted for clinical characteristics are sparse. This study analyzed the impacts of race/ethnicity on outcomes among US patients with COVID-19.

METHODS

This was a retrospective observational study of patients with a confirmed COVID-19 diagnosis in the electronic health record from 01 February 2020 through 14 September 2020. Index encounter site, hospitalization, and mortality were assessed by race/ethnicity (Hispanic, non-Hispanic Black [Black], non-Hispanic White [White], non-Hispanic Asian [Asian], or Other/unknown). Associations between racial/ethnic categories and study outcomes adjusted for patient characteristics were evaluated using logistic regression.

FINDINGS

Among 202,908 patients with confirmed COVID-19, patients from racial/ethnic minority groups were more likely than White patients to be hospitalized on initial presentation (Hispanic: adjusted odds ratio 1·690, 95% CI 1·620-1·763; Black: 1·810, 1·743-1·880; Asian: 1·503, 1·381-1·636) and during follow-up (Hispanic: 1·700, 1·638-1·764; Black: 1·578, 1·526-1·633; Asian: 1·391, 1·288-1·501). Among hospitalized patients, adjusted mortality risk was lower for Black patients (0·881, 0·809-0·959) but higher for Asian patients (1·205, 1·000-1·452).

INTERPRETATION

Racial/ethnic minority patients with COVID-19 had more severe disease on initial presentation than White patients. Increased mortality risk was attenuated by hospitalization among Black patients but not Asian patients, indicating that outcome disparities may be mediated by distinct factors for different groups. In addition to enacting policies to facilitate equitable access to COVID-19-related care, further analyses of disaggregated population-level COVID-19 data are needed.

Intensive remote monitoring versus conventional care in type 1 diabetes: A randomized controlled trial

OBJECTIVE

While frequent contact with diabetes care providers may improve glycemic control among patients with type 1 diabetes (T1D), in-person visits are labor-intensive and costly. This study was conducted to assess the impact of an intensive remote therapy (IRT) intervention for pediatric patients with T1D.

METHODS

Pediatric patients with T1D were randomized to IRT or conventional care (CC) for 6 months. Both cohorts continued routine quarterly clinic visits and uploaded device data; for the IRT cohort, data were reviewed and patients were contacted if regimen adjustments were indicated. Glycated hemoglobin (HbA1c) change from baseline was assessed at 6 and 9 months. Diabetes-related quality of life (QoL), healthcare services utilization, and hypoglycemic events were also tracked.

RESULTS

Among 117 enrollees (60 IRT, 57 CC), mean (SD) 6-month %HbA1c change for IRT vs CC was -0.34 (0.85) (-3.7 mmol/mol) vs -0.05 (0.74) (-0.5 mmol/mol) overall (P = .071); -0.15 (0.67) (1.6 mmol/mol) vs -0.02 (0.66) (0.2 mmol/mol) for ages 8 to 12 (P = .541); and -0.50 (0.95) (-5.5 mmol/mol) vs -0.06 (0.80) (-0.7 mmol/mol) for ages 13 to 17 (P = .056). Diabetes-related QoL increased by 6.5 and 1.3 points for IRT and CC, respectively (P = .062). Three months after intervention cessation, %HbA1c changed minimally among treated children aged 8 to 12 but increased by 0.22 (0.89) (2.4 mmol/mol) among those aged 13 to 17.

CONCLUSIONS

IRT substantially affected diabetes metrics and improved QoL among pediatric patients with T1D. Adolescents experienced a stronger treatment effect, but had difficulty in sustaining improved control after intervention cessation.

Health Care Resource Use and Costs Pre- and Post-Treatment Initiation With Linaclotide: Retrospective Analyses of a U.S. Insured Population

As expected, pharmacy costs increased with the introduction of this new treatment in a market dominated by over-the-counter and generic treatments. On the other hand, outpatient GI-related and irritable bowel disease health care resource use and costs substantially decreased among commercial and Medicare patients following linaclotide treatment initiation.

Can statistical linkage of missing variables reduce bias in treatment effect estimates in comparative effectiveness research studies?

AIM

Missing data, particularly missing variables, can create serious analytic challenges in observational comparative effectiveness research studies. Statistical linkage of datasets is a potential method for incorporating missing variables. Prior studies have focused upon the bias introduced by imperfect linkage.

METHODS

This analysis uses a case study of hepatitis C patients to estimate the net effect of statistical linkage on bias, also accounting for the potential reduction in missing variable bias.

RESULTS

The results show that statistical linkage can reduce bias while also enabling parameter estimates to be obtained for the formerly missing variables.

CONCLUSION

The usefulness of statistical linkage will vary depending upon the strength of the correlations of the missing variables with the treatment variable, as well as the outcome variable of interest.

Characterizing pulmonary hypertension-related hospitalization costs among Medicare Advantage or commercially insured patients with pulmonary arterial hypertension: a retrospective database study

OBJECTIVES

This study assessed pulmonary hypertension (PH)-related hospitalizations, including readmissions, among US patients with pulmonary arterial hypertension (PAH), a rare disease characterized by high morbidity and premature mortality.

STUDY DESIGN

Analysis of claims data (January 1, 2007-April 30, 2011) from adult enrollees with commercial or Medicare Advantage with Part D coverage from a large US health plan.

METHODS

Patients with PAH were identified based on ≥ 1 medical claim with a PH-related diagnostic code (International Classification of Diseases, Ninth Edition, Clinical Modification code 416.0 for primary pulmonary hypertension or 416.8 for other chronic pulmonary heart disease) and ≥ 1 pharmacy claim for a medication indicated for PAH or frequently used in PAH. Data were analyzed for patients with ≥ 1 hospitalization with a primary or secondary diagnostic code of PH. PH-related hospitalizations were defined as those with ≥ 1 PH-related diagnostic code. The principal diagnosis was defined as the diagnosis most frequently in the first-listed position on a hospitalization's facility claims. Total hospitalization costs (inflated to 2011 US$) and length of stay (LOS) were analyzed. A subgroup analysis evaluated readmissions.

RESULTS

Of 4009 enrollees meeting inclusion criteria, 2275 had ≥ 1 PH-related hospitalization during follow-up: 56.9% were female, 59.4% were < 65 years old, and 67.8% had commercial insurance. Mean (SD) costs across all hospitalizations were $46,118 ($135,137) for commercially insured and $16,319 ($30,046) for Medicare Advantage enrollees; LOS was 10.9 (20.4) and 12.8 (21.2) days, respectively. Costs and LOS were higher for admissions with a principal diagnosis of PH compared with other principal diagnoses: $61,922 ($213,596) versus $42,455 ($108,925) and 14.2 (32.3) versus 10.2 (16.4) days, respectively, for the commercially insured, and $19,584 ($29,501) versus $15,904 ($30,097) and 16.7 (25.7) versus 12.3 (20.5) days, respectively, for Medicare Advantage enrollees. Of the 954 patients who experienced ≥ 1 PH-related readmission within the first year after discharge from the initial hospitalization, 483 (50.6%), 246 (25.8%), and 225 (23.6%) patients had 1, 2, and ≥ 3 readmissions, respectively.

CONCLUSIONS

PH-related hospitalizations incur substantial healthcare costs and require long hospital stays for patients with PAH; many are readmitted within 1 year. Improved treatment approaches are needed to reduce PAH disease progression leading to costly and burdensome inpatient stays.

Clinical and economic outcomes associated with amlodipine/renin–angiotensin system blocker combinations

Objectives:

Since treatment regimen type can influence adherence and other outcomes, this study examined adherence, cardiovascular events, and economic outcomes in patients with hypertension treated with fixed-dose combination (FDC) amlodipine/olmesartan (AML/OM), FDC AML/benazepril (AML/BEN), and loose-dose combination AML plus angiotensin II receptor blockers (LDC AML/ARBs).

Methods:

Commercial health plan enrolees aged at least 18 years with index claim(s) for AML/OM, AML/BEN, or LDC AML/ARB were identified. Absence of study drug 6 months pre index, and continuous enrolment for at least 12 months post index were required. Descriptive analyses were executed to make comparisons between treatments, as well as multivariate models adjusting for baseline demographic and clinical characteristics, including propensity for assignment to study drug.

Results:

Descriptive results suggested mean follow-up adherence was higher in the AML/OM cohort [proportion of days covered (PDC) = 0.63] compared with the AML/BEN (PDC = 0.55; p < 0.001) and LDC AML/ARB cohorts (PDC = 0.34; p < 0.001). The proportion of individuals with an incident follow-up cardiovascular event composite was lower in the AML/OM cohort versus the AML/BEN and LDC AML/ARB cohorts (5.94% versus 7.85% and 16.89% respectively). Adjusted Cox models suggested that patients initiated on LDC AML/ARB (hazard ratio 1.35; p < 0.001), but not on AML/BEN, were at greater risk of a follow-up cardiovascular event (composite) compared with AML/OM. Adjusted generalized linear models suggested that mean follow-up per-member-per-month overall costs were higher in the AML/BEN (cost ratio = 1.169; p < 0.001; unadjusted mean cost US$780) and LDC AML/ARB cohorts (cost ratio = 1.286; p < 0.001; unadjusted mean cost US$1394) compared with the AML/OM cohort (unadjusted mean cost US$740).

Conclusion:

The results suggested that treatment with FDC AML/OM was associated with greater likelihood of adherence and lower overall costs than with FDC AML/BEN and LDC AML/ARB, and lower risk of cardiovascular event composite versus LDC AML/ARB.

Costs associated with management of cervical human papillomavirus-related conditions

BACKGROUND

Oncogenic types of human papillomavirus (HPV) have been linked to 99.7% of cervical cancer cases worldwide.

METHODS

This retrospective claims-based analysis was conducted to assess patterns of use and costs associated with diagnostic and treatment procedures for disease attributed to HPV performed before the introduction of HPV vaccination (January 1, 2001-May 31, 2006). Percentages of commercially insured health plan enrollees who underwent each procedure of interest were calculated for each year. Annual costs (combined patient and health plan-paid amounts) were calculated from qualifying medical claims. Descriptive statistics were used to assess trends in procedure rates and costs.

RESULTS

Data for approximately 14.2 million enrollees were obtained. Hysterectomy was the most commonly administered treatment. With the exception of colposcopy with LEEP, all other treatment procedures experienced a decline in rate of use. The most frequently performed diagnostic procedure was colposcopy with endocervical curettage (ECC). With the exception of ECC, rates of diagnostic procedures reached a peak among 20- to 24-year-olds, and followed a downward trend across older groups. Hysterectomy was the most expensive treatment (median $7,383; mean $8,384) per procedure in 2006.

CONCLUSION

Results reveal high rates of use and high-associated costs of diagnostic procedures and treatments related to disease attributed to HPV.

IMPACT

The data presented may be useful in cost-effectiveness analyses and to guide decision makers evaluating how best to optimize prevention strategies.

Abstract 320: Cardiovascular Outcomes Associated with Combination of Renin Angiotensin System Drugs and Amlodipine

To examine cardiovascular outcomes for patients medicated with renin angiotensin system (RAS) drugs/amlodipine in fixed (FDC)- and loose-dose (LDC) combinations. Retrospective claims from a large health plan (data from 01 Mar ‘07 to 31 Dec ’09) were used to identify commercial enrollees aged ≥ 18 yrs with hypertension and index claim for FDC olmesartan/amlodipine (FDC-OA), FDC benazepril/amlodipine (FDC-BA), or LDC angiotensin receptor blocker and amlodipine (LDC-AA). Absence of study drug 6 months prior to index claim, and continuous enrollment for ≥12 months post-index were required. The primary outcome of interest was a cardiovascular event (CVE) composite indicative of heart failure, stroke, myocardial infarction (MI), acute ischemic heart disease excluding MI, and MI/IHD-related surgery. Outcomes were identified on the basis of a primary or secondary ICD-9 diagnosis or procedure codes or CPT codes. To ensure cohort comparability, propensity scores assessing likelihood of assignment to FDC-OA were estimated separately, relative to FDC-BA and LDC-AA, and propensity score quintiles (PSQ) were generated. Cox proportional hazards model of time to first CVE controlling for demographics and baseline characteristics were estimated for each PSQ and results aggregated to compare time to first follow-up event between FDC-OA separately with FDC-BA and LDC-AA. A total of 4,864 individuals were identified on FDC-OA, 12,051 on FDC-BA, and 7,748 on LDC-AA. Mean follow-up duration was 543, 625, and 585 days, respectively. Mean proportion of day covered with therapy was higher in FDC-OA (0.63) cohort compared to FDC-BA (0.55, P <0.001) and LDC-AA cohorts (0.34, P <0.001). The proportion of patients who experienced an incident CVE in the follow-up was lower in the FDC-OA cohort versus FDC-BA and LDC-AA cohorts (5.94% vs 7.85%, and 16.85%, respectively). Adjusted Cox models suggested that patients initiated on LDC-AA (Hazard ratio [HR] =1.35, p<.001) but not FDC-BA (HR=1.14, p=0.085) were at greater risk of having a CVE compared to FDC-OA. Hazard ratios for first CVE in every quintile were greater for the LDC-AA compared to FDC-OA. In this large, managed care population, patients initiated with FDC-OA had a lower risk of cardiovascular events versus those initiated on LDC-AA, which may be attributed to the greater adherence associated with FDC therapy.

Factors related to immunosuppressant medication adherence in renal transplant recipients

Non-adherence to immunosuppressant medications (ISM) is a significant issue for transplant recipients. This study examines factors influencing ISM adherence in renal transplant recipients (RTRs). Patient-reported data were collected through a cross-sectional survey including use of ISMs, adherence behaviors, perceived adherence barriers, beliefs and attitudes toward ISMs, and patient life satisfaction. Logistic regression was conducted to examine how RTRs' beliefs about use of ISMs, life satisfaction, and ISM adherence barriers were related to adherence. A total of 512 adult commercial insurance enrollees following renal transplantation were included in the analysis. One hundred and seventy-seven RTRs were non-adherent (34.5%); the most frequently cited reason was forgetfulness. RTRs aged 18-29 yr were more likely to be non-adherent than recipients 46-64 yr old (p ≤ 0.001). Non-adherent RTRs had greater adherence barriers than adherent RTRs (p < 0.001). Adherent RTRs believed their ISMs were more necessary than non-adherent RTRs (p < 0.001), while non-adherent RTRs had greater concerns about taking ISMs (p = 0.009) and believed they had less control over their lives than adherent RTRs (p < 0.001). Non-adherent RTRs had lower life satisfaction (p < 0.001). Non-adherence is significantly associated with patients' beliefs about ISMs, perceived barriers, and lower life satisfaction. Strategies to increase ISM adherence are discussed.

Long-term clinical and economic outcomes associated with angiotensin II receptor blocker use in hypertensive patients

OBJECTIVE

To examine clinical and economic outcomes associated with angiotensin II receptor blockers (ARB).

METHODS

Retrospective claims data were analyzed for hypertensive adults with ≥1 year follow-up from first ARB claim. Subjects were stratified into four cohorts: olmesartan (OM); valsartan (VAL); losartan (LOS); and irbesartan (IRB), which represented the full sample. Analyses were also conducted with the "limited sample," which excluded subjects with pre-existing conditions in the period before first ARB. Time to follow-up cardiac event was modeled using Cox proportional hazards regression; follow-up healthcare resource utilization and costs were examined using generalized linear models.

RESULTS

The full and limited samples consisted of 118,700 and 65,579 subjects, respectively. Mean follow-up ranged from 861 to 933 days. Baseline characteristics including the Quan-Charlson comorbidity score differed by cohort. In both the full and limited samples, respectively, multivariate models predicted a higher adjusted risk of follow-up cardiac event for VAL cohort (hazard ratio [HR] = 1.261 and 1.242, p < 0.001), LOS cohort (HR = 1.307 and 1.178, p < 0.01), and IRB cohort (HR = 1.222 and 1.179, p ≤ 0.016) compared to OM cohort. Adjusted risk (full sample) of follow-up ambulatory and inpatient visits (all-cause and hypertension-attributable) was higher in VAL, LOS, and IRB cohorts compared to OM. Adjusted risk (limited sample) of follow-up ambulatory visits (all-cause and hypertension-attributable) was greater for VAL, LOS and IRB cohorts relative to OM, but inpatient visit risk was greater only in VAL and LOS cohorts. Compared to the OM cohort, follow-up all-cause adjusted healthcare costs (limited sample) were higher in VAL (cost ratio [CR] = 1.067, p = 0.001) and IRB cohorts (CR = 1.062, p = 0.045).

CONCLUSIONS

In this large national US health plan, treatment with OM was associated with lower risk of cardiac events and lower healthcare resource utilization and costs versus VAL, LOS, and IRB over a mean follow-up of 2.5 years. Association, rather than causality, to cardiac outcomes may be inferred from these observational claims data.

Physician perception of patient adherence compared to patient adherence of osteoporosis medications from pharmacy claims

OBJECTIVE

This study explored physicians' perceptions of patient adherence to medications compared with patient adherence derived by administrative data in the treatment of osteoporosis.

RESEARCH DESIGN AND METHODS

A study involving written questionnaires from prescribers treating patients with postmenopausal osteoporosis (PMO) compared the questionnaire responses to pharmacy claims of these prescribers' patients' refill patterns. Approximately 2000 physicians from a large US health plan were faxed or mailed a survey. Data from the physician survey were merged with administrative claims data of the participating physicians' patients.

RESULTS

A total of 412 physicians (21.8%) responded. Although a low response rate, there were no significant demographic differences between participating and non-participating physicians. Surveyed physicians reported that 66% of their patients had private/commercial coverage and over 60% reported seeing their PMO patients annually. Overall, physicians estimated that 69.2% of patients were adherent 80% of the time after 12 months of therapy. Yet, pharmacy claims data for those physicians' patients indicated 48.7% of patients were adherent (defined as having an MPR of >or=80%) after 12 months of therapy. Physicians overestimated their patients' adherence regardless of medication class and across physician specialties. Regression modeling revealed that physicians who have been in practice longer estimated fewer patients as adherent, whereas those who prescribe more PMO treatments estimate a greater number of patients as adherent. Providers cited side effects and affordability of medication as the most frequent reasons for non-adherence.

CONCLUSIONS

Physicians overestimate patient adherence to PMO therapies. Improving physician awareness of medication non-adherence to PMO therapies may facilitate physician-patient dialogue, with the aim of identifying patient-centered reasons for non-adherence. These discussions are important because patients with poorer adherence have a higher risk of fracture. Future research should focus on reasons for patient non-adherence to osteoporosis regimens and intervention strategies that improve communication between the provider and patient. Findings must be considered within the limitations of this claims database analysis. Some degree of incomplete or incorrect coding may exist, and the presence of a claim for a filled prescription does not indicate that the medication was consumed or taken as prescribed. Patients included in the study are not necessarily representative of all patients being treated for osteoporosis.