The presence of statistical outliers is a shared concern in research. If ignored or improperly handled, outliers have the potential to distort the estimate of the parameter of interest and thus compromise the generalizability of research findings. A variety of statistical techniques are available to assist researchers with the identification and management of outlier cases. The purpose of this paper is to provide a conceptual overview of univariate outliers with special focus on common techniques used to detect and manage univariate outliers. Specifically, this paper discusses the use of histograms, boxplots, interquartile range, and z-score analysis as common univariate outlier identification techniques. The paper also discusses the outlier management techniques of deletion, substitution, and transformation.

BackgroundThe 2016 Canadian Triage and Acuity Scale (CTAS) updates introduced frailty screening within triage to more accurately code frail patients who may deteriorate waiting for care. The relationship between triage acuity and frailty is not well understood, but may help inform which supplemental geriatric assessments are beneficial to support care in the emergency department (ED). Our objectives were to investigate the relationship between triage acuity and frailty, and to compare their associations with a series of patient outcomes.MethodsWe conducted a secondary analysis of the Canadian cohort from a multinational prospective study. Data were collected on ED patients 75 years of age and older from eight ED sites across Canada between November 2009 and April 2012. Triage acuity was assigned using the CTAS, whereas frailty was measured using an ED frailty index. Spearman rank and binary logistic regression were used to examine associations.ResultsA total of 2,153 ED patients were analyzed. No association was found between the CTAS and ED frailty index scores assigned to patients (r = .001; p = 0.99). The ED frailty index was associated with hospital admission (odds ratio [OR] = 1.5; 95% confidence interval [CI] = 1.4–1.6), hospital length of stay (OR = 1.4; 95% CI = 1.2–1.6), future hospitalization (OR = 1.1; 95% CI = 1.05–1.2), and ED recidivism (OR = 1.1; 95% CI = 1.04–1.2). The CTAS was associated with hospital admission (e.g., CTAS 2 v. 5; OR = 6; 95% CI = 3.3–11.4).ConclusionOur findings demonstrate that frailty and triage acuity are independent but complementary measures. EDs may benefit from comprehensive frailty screening post-triage, as frailty and its associated geriatric syndromes drive outcomes separate from traditional measures of acuity.

No abstract

BACKGROUND Home‐based primary care has been associated with reductions in hospital use among homebound older adults, but population‐based studies on the general home visit patterns of primary care physicians are lacking. OBJECTIVE We examined the association between the provision of home visits by primary care physicians and subsequent use of hospital‐based care among their older adult patients with extensive functional impairments. DESIGN Population‐based retrospective cohort study. SETTING The setting was Ontario, Canada, from October 2014 to September 2016. PARTICIPANTS Older adults (aged ≥65 years) with extensive functional impairments receiving publicly funded home care. MEASUREMENTS We measured the provision of home visits by a patientʼs most responsible primary care physician during the year before a comprehensive home care assessment. Physician home visit patterns were measured as the proportion of the total outpatient visits in a year that were home visits, categorized with quartiles. Multivariable, multilevel negative binomial regression models examined the associations between physician‐level home visit provision and patient emergency department visits and hospital admissions over the 6 months following the home care assessment. RESULTS There were 49,613 patients in the cohort who were linked to 8,096 unique primary care physicians. A total of 69.1% of physicians provided at least one home visit in a year, with the median proportion of home visits to total visits ranging from 0.057% to 3.19% across quartiles. Patients whose physicians were in the highest home visit provision quartile had lower rates of emergency department visits (incidence rate ratio [IRR] = 0.93; 95% confidence interval [CI] = 0.90–0.96) and hospital admissions (IRR = 0.89; 95% CI = 0.85–0.93) compared with patients whose physician did not do home visits. CONCLUSION Home care patients with extensive functional impairments whose physicians provided higher levels of home visits had fewer emergency department visits and hospital admissions. Expanding home visits by primary care physicians could reduce hospital use by older adults living with functional impairments in the community.

No abstract

Background Increasing hospitalization rates present unique challenges to manage limited inpatient bed capacity and services. Transport by paramedics to the emergency department (ED) may influence hospital admission decisions independent of patient need/acuity, though this relationship has not been established. We examined whether mode of transportation to the ED was independently associated with hospital admission. Methods We conducted a retrospective cohort study using the National Ambulatory Care Reporting System (NACRS) from April 1, 2015 to March 31, 2020 in Ontario, Canada. We included all adult patients (≥18 years) who received a triage score in the ED and presented via paramedic transport or self-referral (walk-in). Multivariable binary logistic regression was used to determine the association of mode of transportation between hospital admission, after adjusting for important patient and visit characteristics. Results During the study period, 21,764,640 ED visits were eligible for study inclusion. Approximately one-fifth (18.5%) of all ED visits were transported by paramedics. All-cause hospital admission incidence was greater when transported by paramedics (35.0% vs. 7.5%) and with each decreasing Canadian Triage and Acuity Scale level. Paramedic transport was independently associated with hospital admission (OR = 3.76; 95%CI = 3.74–3.77), in addition to higher medical acuity, older age, male sex, greater than two comorbidities, treatment in an urban setting and discharge diagnoses specific to the circulatory or digestive systems. Conclusions Transport by paramedics to an ED was independently associated with hospital admission as the disposition outcome, when compared against self-referred visits. Our findings highlight patient and visit characteristics associated with hospital admission, and can be used to inform proactive healthcare strategizing for in-patient bed management.

No abstract

No abstract