Little is known about whether diabetes increases the risk of COVID-19 infection and whether measures of diabetes severity are related to COVID-19 outcomes. Investigate diabetes severity measures as potential risk factors for COVID-19 infection and COVID-19 outcomes. In integrated healthcare systems in Colorado, Oregon, and Washington, we identified a cohort of adults on February 29, 2020 (n = 1,086,918) and conducted follow-up through February 28, 2021. Electronic health data and death certificates were used to identify markers of diabetes severity, covariates, and outcomes. Outcomes were COVID-19 infection (positive nucleic acid antigen test, COVID-19 hospitalization, or COVID-19 death) and severe COVID-19 (invasive mechanical ventilation or COVID-19 death). Individuals with diabetes (n = 142,340) and categories of diabetes severity measures were compared with a referent group with no diabetes (n = 944,578), adjusting for demographic variables, neighborhood deprivation index, body mass index, and comorbidities. Of 30,935 patients with COVID-19 infection, 996 met the criteria for severe COVID-19. Type 1 (odds ratio [OR] 1.41, 95% CI 1.27–1.57) and type 2 diabetes (OR 1.27, 95% CI 1.23–1.31) were associated with increased risk of COVID-19 infection. Insulin treatment was associated with greater COVID-19 infection risk (OR 1.43, 95% CI 1.34–1.52) than treatment with non-insulin drugs (OR 1.26, 95% 1.20–1.33) or no treatment (OR 1.24; 1.18–1.29). The relationship between glycemic control and COVID-19 infection risk was dose-dependent: from an OR of 1.21 (95% CI 1.15–1.26) for hemoglobin A1c (HbA1c) < 7% to an OR of 1.62 (95% CI 1.51–1.75) for HbA1c ≥ 9%. Risk factors for severe COVID-19 were type 1 diabetes (OR 2.87; 95% CI 1.99–4.15), type 2 diabetes (OR 1.80; 95% CI 1.55–2.09), insulin treatment (OR 2.65; 95% CI 2.13–3.28), and HbA1c ≥ 9% (OR 2.61; 95% CI 1.94–3.52). Diabetes and greater diabetes severity were associated with increased risks of COVID-19 infection and worse COVID-19 outcomes.

Results.Of 35,886 outpatients included in the analysis, 9,700 (27%) had a provider launch their "Patient Values Tab."Of 10,055 inpatients included in the analysis, 4,517 (45%) had a provider launch their "Patient Values Tab." Patterns of adoption varied by provider role, with 67% of chaplains, 77% of social workers, 49% of advanced practice providers, 32% of nurses, and 27% of attending physicians launching the "Patient Values Tab" for at least one patient; this was consistent across outpatient and inpatient cohorts. Conclusion. Patterns of adoption of an innovative"Patient Values Tab" EHR feature varied across clinical contexts and provider roles.Ongoing research is exploring factors that impede/facilitate adoption at the level of the patient, provider, and clinical context to inform strategies to enhance/sustain adoption.

Background Patients, family members, and clinicians express concerns about potential adverse drug withdrawal events (ADWEs) following medication discontinuation or fears of upsetting a stable medical equilibrium as key barriers to deprescribing. Currently, there are limited methods to pragmatically assess the safety of deprescribing and ascertain ADWEs. We report the methods and results of safety monitoring for the OPTIMIZE trial of deprescribing education for patients, family members, and clinicians. Methods This was a pragmatic cluster randomized trial with multivariable Poisson regression comparing outcome rates between study arms. We conducted clinical record review and adjudication of sampled records to assess potential causal relationships between medication discontinuation and outcomes. This study included adults aged 65+ with dementia or mild cognitive impairment, one or more additional chronic conditions, and prescribed 5+ chronic medications. The intervention included an educational brochure on deprescribing that was mailed to patients prior to primary care visits, a clinician notification about individual brochure mailings, and an educational tip sheets was provided monthly to primary care clinicians. The outcomes of the safety monitoring were rates of hospitalizations and mortality during the 4 months following brochure mailings and results of record review and adjudication. The adjudication process was conducted throughout the trial and included classifications: likely, possibly, and unlikely. Results There was a total of 3012 (1433 intervention and 1579 control) participants. There were 420 total hospitalizations involving 269 (18.8%) people in the intervention versus 517 total hospitalizations involving 317 (20.1%) people in the control groups. Adjusted risk ratios comparing intervention to control groups were 0.92 [95% confidence interval (CI) 0.72, 1.16] for hospitalization and 1.19 (95% CI 0.67, 2.11) for mortality. Both groups had zero deaths “likely” attributed to a medication change prior to the event. A total of 3 out of 30 (10%) intervention group hospitalizations and 7 out of 35 (20%) control group hospitalizations were considered “likely” due to a medication change. Conclusions Population-based deprescribing education is safe in the older adult population with cognitive impairment in our study. Pragmatic methods for safety monitoring are needed to further inform deprescribing interventions. Trial Registration NCT03984396. Registered on 13 June 2019

Persons living with dementia are likely to benefit from interventions to deprescribe potentially inappropriate medications (PIM). PIM are medications for which risks of use outweigh benefits or safer alternatives exist. Variation in PIM definitions limits comparison of outcomes across deprescribing studies.1-3 To test this effect, we compared the effect of using two different PIM lists in the OPTIMIZE (Optimal Medication Management in Alzheimer Disease and Dementia) trial. OPTIMIZE was a pragmatic, cluster randomized trial of deprescribing education versus usual care for patients with cognitive impairment and their families. It was conducted from April 1, 2019, to March 31, 2020, at primary care clinics within Kaiser Permanente Colorado (KPCO), a not-for-profit integrated healthcare system.4, 5 The KPCO and Johns Hopkins University institutional review boards approved the study. Eligibility criteria included mild cognitive impairment or dementia, ≥1 additional chronic conditions, and ≥5 chronic medications. The intervention was designed to raise awareness of deprescribing and did not specifically focus on PIM. The primary PIM outcome was the percentage of individuals prescribed ≥1 PIM at 6 months. We compared this outcome using the original PIM list (Supplementary Table S1) to a revised PIM list. The original list was based in part on the Beers list of drugs to avoid for individuals with cognitive impairment5, 6 plus opioids. The revised PIM list included all medications from the 2019 Revised Beers list7 and additional medications that might be broadly targeted for deprescribing (e.g., proton pump inhibitors [PPIs]) and was edited for prescription utilization at KPCO. This resulted in 56 medications being removed from the original list, 105 medications kept, and 71 medications added. Analysis was performed on an intention-to-treat basis. Odds of taking ≥1 PIM were modeled using logistic regression accounting for baseline PIM use. Models also included self-reported race/ethnicity, age, and sex. Statistical analysis was conducted with SAS, version 9.4. All p values were from two-sided tests; results were deemed statistically significant at p < 0.05. A sensitivity analysis examined whether effect estimates changed when PPIs were removed from the revised list. Of 3012 patients (Supplementary Table S2), the percentage of patients taking PIM at baseline was 393 of 1433 (27.4%) in the intervention group and 435 of 1579 (27.6%) in the control group using the original PIM list. Using the revised list, the percentages were much higher: 908 of 1433 (63.4%) took PIM in the intervention group and 1054 of 1579 (66.8%) in the control group. In the sensitivity analysis excluding PPIs, 694 of 1433 (48.4%) took PIM in the intervention group and 804 of 1579 (50.9%) in the control group. The percentages of patients taking PIM at 6 and 12 months are in Figure 1. In logistic regressions, the odds of being on PIM were similar between intervention and control groups at 6 months using the original list (Table 1). By contrast, using the revised list, the odds of being on PIM were significantly lower in the intervention versus control groups. In the sensitivity analysis excluding PPIs, the estimate was no longer significant. At 12 months, the odds of being on PIM were similar in the intervention versus control groups using both lists and in the sensitivity analysis excluding PPIs (Table 1). In this secondary analysis, we found that the absolute number of patients taking PIM was much higher using the revised PIM list. The effectiveness of the intervention for the PIM outcome depended on how PIM were defined. These results have implications for the selection of medication appropriateness measures for deprescribing interventions. Deprescribing trials should specify which medications are included in their PIM definitions and why. The lack of standardized definitions of PIM for specific populations is a hindrance to deprescribing evidence generation and a limitation to informing implementation efforts.3, 8 In our sensitivity analysis excluding PPIs, the intervention effect was no longer significant at 6 months. This suggests that PPIs were influential in increasing the proportion of people identified as taking PIM and may be a high-yield target for deprescribing interventions.9 Study limitations include lack of statistical power to identify which drug classes were most likely to be deprescribed. We may have been more likely to detect a difference in the percentage of individuals taking PPIs because they were commonly prescribed at baseline. Additionally, we were unable to determine if continuation of a PIM was consistent with patient goals or reflected a failed withdrawal trial. In summary, our findings suggest that to evaluate the relative benefits of deprescribing interventions, future research will need PIM outcome measure definitions that are comparable across studies. At 6 months follow-up, using the original PIM list, the percentage of patients taking PIM was 370 of 1433 (25.8%) in the intervention group and 417 of 1579 (26.4%) in the control group. Using the revised PIM list, 864 of 1433 (60.3%) were taking PIM in the intervention group and 1043 of 1579 (66.1%) in the control group. In the sensitivity analysis excluding PPIs from the revised list, 663 of 1433 (46.3%) were taking PIM in the intervention group and 782 of 1579 (49.5%) in the control group. At 12 months follow-up, using the original PIM list, the percentage of patients taking PIM was 351 of 1433 (24.5%) in the intervention group and 396 of 1579 (25.1%) in the control group. Using the revised PIM list, 865 of 1433 (60.4%) were taking PIM in the intervention group and 1001 of 1579 (63.4%) in the control group. In the sensitivity analysis excluding PPIs from the revised list, 662 of 1433 (46.2%) were taking PIM in the intervention group and 755 of 1579 (47.8%) in the control group. Ariel R. Green contributed to study concept and design, interpretation of data and drafting and drafting and revision of the manuscript. Elizabeth A. Bayliss and Cynthia M. Boyd contributed to study concept and design, acquisition of subjects and/or data, analysis and interpretation of data and critical review of manuscript. Linda A. Weffald, Melanie L. Drace, and Jonathan D. Norton contributed to interpretation of data and critical review of manuscript. John D. Powers contributed to study concept and design, acquisition of subjects and/or data, analysis and interpretation of data and critical review of manuscript. We thank Susan M. Shetterly for performing primary statistical analyses on the OPTIMIZE trial. Cynthia M. Boyd reported receiving royalties from UpToDate for writing a chapter on multi-morbidity and honoraria from Dynamed for reviewing a chapter on falls outside the submitted work. No other disclosures were reported. None. The study was funded by the National Institute on Aging, Grant/Award Number: R33-AG057289. Supplementary Table S1. Potentially inappropriate medication lists. Supplementary Table S2. Characteristics of the study population. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.

Abstract Examining end-of-life (EOL) patient outcomes associated with patient portal use will help identify benefits and barriers of portal use during EOL care. The primary objective of this retrospective cohort study was to assess the association between EOL outcomes and patient portal use among a cohort (N=6,517) of deceased patients (18+ years old) in the last 12 months of life within Kaiser Permanente Colorado (KPCO). Outcomes including 1) number of hospitalizations in the last 30 days of life, 2) having documented advance directive before death, and 3) dying in hospice were investigated by levels of patient portal use (no use, non-active use, active use without provider, and active use with a provider) for the first six months of the last year of life. A generalized linear model with Poisson distribution, a log link and an offset to account for person-days of follow-up was used to estimate rates of hospitalizations in the last 30 days. Logistic regression was used to estimate advance directive and death in hospice outcomes. Patient portal use appeared to be associated with beneficial EOL outcomes such as advance care planning (P = 0.0226) and dying in hospice (P = 0.0070). Hospitalizations were greater with higher levels of patient portal use (P = 0.0492), suggesting that patients/caregivers may have used patient portals for disease related information and accessing health services during times of need. More research is needed to leverage patient portal use to intentionally engage, educate and guide patients with serious illness and approaching EOL.

INTRODUCTION:Colorectal cancer (CRC) incidence and mortality can be reduced by effective screening and/or treatment. However, the influence of health care systems on disparities among insured patients is largely unexplored.METHODS:To evaluate insured patients with CRC diagnosed between 2010 and 2014 across 6 diverse US health care systems in the Patient-Centered Outcomes Research Institute (PCORI) Patient Outcomes Research To Advance Learning (PORTAL) CRC cohort, we contrasted CRC stage; CRC mortality; all-cause mortality; and influences of demographics, stage, comorbidities, and treatment between health systems.RESULTS:Among 16,211 patients with CRC, there were significant differences between health care systems in CRC stage at diagnosis, CRC-specific mortality, and all-cause mortality. The unadjusted risk of CRC mortality varied from 27% lower to 21% higher than the reference system (hazard ratio [HR] = 0.73, 95% confidence interval = 0.66-0.80 to HR = 1.21, 95% confidence interval = 1.05-1.40; p < 0.01 across systems). Significant differences persisted after adjustment for demographics and comorbidities (p < 0.01); however, adjustment for stage eliminated significant differences (p = 0.24). All-cause mortality among patients with CRC differed approximately 30% between health care systems (HR = 0.89-1.17; p < 0.01). Adjustment for age eliminated significant differences (p = 0.48).DISCUSSION:Differences in CRC survival between health care systems were largely explained by stage at diagnosis, not demographics, comorbidity, or treatment. Given that stage is strongly related to early detection, these results suggest that variation in CRC screening systems represents a modifiable systems-level factor for reducing disparities in CRC survival.

AbstractBackgroundNumerous studies have examined melanoma incidence and survival, but studies on melanoma recurrence are limited. We examined melanoma incidence, recurrence, and mortality among members of Kaiser Permanente Colorado (KPCO) between January 1, 2000 and December 31, 2015.MethodsAge‐adjusted incidence rates were computed to examine trends among KPCO members aged 21 years and older. Cox proportional hazards models were used to examine factors associated with recurrence and mortality.ResultsOur cohort included 1931 cases of invasive melanoma. Incidence rates increased over time and were higher than SEER rates; however, the increase was limited to early stage disease. In multivariable models, stage at initial diagnosis, gender, and age were associated with melanoma recurrence. Men were more likely to have a recurrence than women (adjusted hazard ratio [HR]: 1.70, 95% confidence interval [CI]: 1.19‐2.43), and for each decade of increasing age, the adjusted HR = 1.20 (95% CI: 1.06‐1.37). Factors associated with all‐cause mortality included stage (HR = 12.87, 95% CI: 6.63‐24.99, for stage IV vs stage I), male gender (HR = 1.42, 95% CI: 1.12‐1.79), older age at diagnosis, lower socioeconomic status, and comorbidity index. For melanoma‐specific mortality, results were similar, with one exception: age was not associated with melanoma‐specific death (HR = 1.09, 95% CI: 0.94‐1.25, P = 0.253).ConclusionsData derived from an insured patient population, such as KPCO, have the potential to enhance our understanding of emerging trends in melanoma. This is the first population‐based study in the United States to examine patient characteristics associated with risk of recurrence. Men have an increased risk of both recurrence and death, and thus may benefit from more intensive follow‐up than women.

566 Background: While overall incidence and mortality of colorectal cancer (CRC) has declined, incidence and mortality are increasing in those < 50 years old (early-onset CRC). Our objective was to better understand early-onset CRC by comparing tumor characteristics and initial treatment type between early-onset and normal/late-onset CRC. Methods: We used health system and national tumor registries to identify patients diagnosed with adenocarcinoma of the colon or rectum from 2010-2014 at 6 US integrated health systems in the Patient Outcomes To Advance Learning (PORTAL) network. Tumor registry data included: age at diagnosis, stage, grade, anatomic site, histology, number of lymph nodes examined, and receipt of initial systemic therapy (chemotherapy or immunotherapy). Demographics and other patient characteristics were obtained from the EHR. We used logistic regression to calculate adjusted odds ratios (ORs) and 95% confidence intervals (CIs) comparing the distribution of tumor characteristics and treatment patterns in early-onset ( < 50 years old) vs. normal/late-onset CRC. Results: There were 1,424 early-onset and 10,810 normal/late onset CRC cases in our analyses. Compared to normal/late onset cases, patients with early-onset CRC were more likely to be Hispanic, obese, never smokers, and to have Charlson comorbidity scores < 3. After adjustment for patient characteristics, compared to normal/late onset CRC, early-onset CRC was associated with more advanced stage disease (OR for stage 4 vs. stage 1 = 2.8, CI: 2.4-3.4), high grade histology (OR for poorly differentiated/undifferentiated vs. well/moderately differentiated = 1.2, CI: 1.1-1.5), signet ring histology (OR for signet ring vs. non-mucinous adenocarcinoma = 1.7, CI: 1.1, 2.6), and rectal (OR for rectum vs. cecum = 2.4, CI: 1.9-2.9) or left colon location (OR for left colon vs. cecum = 2.2, CI: 1.8-2.8). After adjustment for patient and tumor characteristics, early-onset patients were more likely than normal/late onset patients to have > 12 lymph nodes examined (OR = 1.6, CI: 1.4-1.8) and to receive systemic therapy (OR = 2.8, CI: 2.4, 3.4). Conclusions: Early-onset CRC is associated with aggressive tumor characteristics, distal location, and systemic therapy use.

BACKGROUND:For privacy and practical reasons, it is sometimes necessary to minimize sharing of individual-level information in multisite studies. However, individual-level information is often needed to perform more rigorous statistical analysis.OBJECTIVES:To compare empirically 3 analytic methods for multisite studies that only require sharing of summary-level information to perform statistical analysis that have traditionally required access to detailed individual-level data from each site.RESEARCH DESIGN, SUBJECTS, AND MEASURES:We analyzed data from a 7-site study of bariatric surgery outcomes within the Scalable Partnering Network. We compared the long-term risk of rehospitalization between adjustable gastric banding and Roux-en-y gastric bypass procedures using a stratified analysis of propensity score (PS)-defined strata, a case-centered analysis of risk set data, and a meta-analysis of site-specific effect estimates. Their results were compared with the result from a pooled individual-level data analysis.RESULTS:The study included 1327 events (18.1%) among 7342 patients. The adjusted hazard ratio was 0.71 (95% CI, 0.59, 0.84) comparing adjustable gastric banding with Roux-en-y gastric bypass in the individual-level data analysis. The corresponding effect estimate was 0.70 (0.59, 0.83) in the PS-stratified analysis, 0.71 (0.59, 0.84) in the case-centered analysis, and 0.71 (0.60, 0.84) in both the fixed-effect and random-effects meta-analysis.CONCLUSIONS:In this empirical study, PS-stratified analysis, case-centered analysis, and meta-analysis produced results that are identical or highly comparable with the result from a pooled individual-level data analysis. These methods have the potential to be viable analytic alternatives when sharing of individual-level information is not feasible or not preferred in multisite studies.