No full-text available
To read the full-text of this research,
you can request a copy directly from the authors.
Adapting a clinical comorbidity index for use with ICD-9-CM administrative data
... To define medications, we used Anatomical Therapeutic Chemical classification codes in the PharmaNet database for antidepressant (N06A) and anxiolytics (N05B). 24 Covariate ascertainment. We assessed covariates for each yearly interval in the five years before and five years after index date. ...
... Covariates that we included were age, sex, neighborhood income quintile, residence (urban or rural), and comorbidities as measured using the Charlson-Romano Comorbidity Index. 24 Statistical analysis. We assessed descriptive statistics for IA and IA-free controls in the year before the index date. ...
Objective
The study objective was to describe patterns of depression and anxiety health care use before and after diagnosis among patients with inflammatory arthritis (IA), namely, ankylosing spondylitis, psoriatic arthritis, and rheumatoid arthritis.
Methods
We used population‐based linked administrative health data from British Columbia, Canada, to build a cohort of individuals (≥18 years) with incident IA and individuals without IA (“IA‐free controls”) matched on age and sex. We computed the proportion of individuals with IA and controls who had one or more depression or one or more anxiety health care encounters and the use of one or more antidepressants or one or more anxiolytics in each yearly interval five years before and after IA diagnosis. We used multivariable logistic regression models to evaluate the association between IA status and aforementioned depression and anxiety health care use outcomes in each yearly interval.
Results
A total of 80,238 individuals with IA (62.9% female; mean ± SD age 56.2 ± 16.7 years) and 80,238 IA‐free controls (62.9% female; mean ± SD age 56.2 ± 16.6 years) were identified between January 1, 2001, and March 31, 2018. Individuals with IA had significantly increased odds of depression and anxiety health care encounters and dispensation of antidepressants and anxiolytics for each yearly interval before and after diagnosis. Adjusted odds ratios (ORs) were highest in the year immediately before (one or more depression visits: adjusted OR 1.61, 95% confidence interval [CI] 1.55–1.66; one or more anxiolytics: adjusted OR 1.71, 95% CI 1.66–1.77) or after (one or more antidepressants: adjusted OR 1.95, 95% CI 1.89–2.00) IA diagnosis.
Conclusion
Findings suggest a role for depression and anxiety in characterizing the IA prodrome period and generate hypotheses regarding overlapping biopsychosocial processes that link IA and mental health comorbidities.
image
... In the health services literature, widely used risk adjustment models based on coded comorbidities include the Charlson comorbidity index (CCI) [14], the Charlson/Deyo index [15], the Charlson/D'Hoore index [16], the Charlson/Romano index [17], and the Elixhauser index (EI) [18]. Those indices are based on a standard system for coding diagnoses, the International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) codes, from administrative health data of hospitalization or outpatient visit. ...
... In addition, due to the likelihood that coding of the tailing digits in ICD-9 codes may lead to inconsistencies, therefore, D'Hoore et al. had declared that the Charlson/ D'Hoore index was a more reliable comorbidity measure. The Charlson/Romano index [17], originally termed as the Dartmouth-Manitoba CCI, was firstly created by Roos et al. in 1989 and subsequently modified by Romano and colleagues in 1993. Compared with the Deyo CCI, the Romano CCI contains more ICD-9-CM codes. ...
Objectives
This study assessed risk adjustment performance of six comorbidity indices in two categories of comorbidity measures: diagnosis-based comorbidity indices and medication-based ones in patients with chronic obstructive pulmonary disease (COPD).
Methods
This was a population–based retrospective cohort study. Data used in this study were sourced from the Taiwan National Health Insurance Research Database. The study population comprised all patients who were hospitalized due to COPD for the first time in the target year of 2012. Each qualified patient was individually followed for one year starting from the index date to assess two outcomes of interest, medical expenditures within one year after discharge and in-hospital mortality of patients. To assess how well the added comorbidity measures would improve the fitted model, we calculated the log-likelihood ratio statistic G². Subsequently, we compared risk adjustment performance of the comorbidity indices by using the Harrell c-statistic measure derived from multiple logistic regression models.
Results
Analytical results demonstrated that that comorbidity measures were significant predictors of medical expenditures and mortality of COPD patients. Specifically, in the category of diagnosis-based comorbidity indices the Elixhauser index was superior to other indices, while the RxRisk-V index was a stronger predictor in the framework of medication-based codes, for gauging both medical expenditures and in-hospital mortality by utilizing information from the index hospitalization only as well as the index and prior hospitalizations.
Conclusions
In conclusion, this work has ascertained that comorbidity indices are significant predictors of medical expenditures and mortality of COPD patients. Based on the study findings, we propose that when designing the payment schemes for patients with chronic diseases, the health authority should make adjustments in accordance with the burden of health care caused by comorbid conditions.
... Socioeconomic status was defined as the patient's neighbourhood median household income quintile from the Statistics Canada Census. The Romano adaptation of the Charlson comorbidity index was calculated based on both inpatient and outpatient encounters in the 3 years prior to cohort entry (excluding RA encounters from the index computation) (19). Additionally, the Johns Hopkins ACG ® System was used to identify the number of aggregated diagnosis groups (ADGs) to characterize the health status of an individual based on hospital discharge abstracts and physician claims data in the 3 years prior to cohort entry. ...
... Univariate and multivariable logistic regression were used to evaluate potential predictors of remaining under rheumatologist care throughout the entire first 5 years following index date. Variables included in the adjusted model included the following: patient age (in 10-year intervals), sex, income quintile, year of diagnosis, Charlson score (Romano adaptation) (19), number of ADGs (categorized as 5-9 or ≥10), rurality, FTE rheumatologist supply per population (benchmark of 1/75,000 adults), distance (per 10 km) to the rheumatologist, living at a remote distance to the rheumatologist, and rheumatologist sex and age (per 10 years). All regressions were adjusted for clustering within the first rheumatologist seen by the patient. ...
Objective:
The study objective was to assess adherence to system-level performance measures measuring retention in rheumatology care and disease modifying anti-rheumatic drug (DMARD) treatment in rheumatoid arthritis (RA).
Methods:
We used a validated health administrative data case definition to identify individuals with RA in Ontario, Canada, between 2002 and 2014 who had at least 5 years of potential follow-up prior to 2019. During the first 5 years following diagnosis, we assessed whether patients were seen by a rheumatologist yearly and the proportion dispensed a DMARD yearly (in those aged ≥66 for whom medication data were available). Multivariable logistic regression analyses were used to estimate the odds of remaining under rheumatologist care.
Results:
The cohort included 50,883 patients with RA (26.1% aged 66 years and older). Over half (57.7%) saw a rheumatologist yearly in all 5 years of follow-up. Sharp declines in the percentage of patients with an annual visit were observed in each subsequent year after diagnosis, although a linear trend to improved retention in rheumatology care was seen over the study period (P < 0.0001). For individuals aged 66 years or older (n = 13,293), 82.1% under rheumatologist care during all 5 years after diagnosis were dispensed a DMARD annually compared with 31.0% of those not retained under rheumatology care. Older age, male sex, lower socioeconomic status, higher comorbidity score, and having an older rheumatologist decreased the odds of remaining under rheumatology care.
Conclusion:
System-level improvement initiatives should focus on maintaining ongoing access to rheumatology specialty care. Further investigation into causes of loss to rheumatology follow-up is needed.
... For this study, the covariates we evaluated were age, sex, neighbourhood income quintile, residence (urban/rural), and the Charlson-Romano Comorbidity Index to represent comorbidities [28]. ...
Objectives
Describe patterns of pharmacotherapy and psychological treatment and evaluate receipt of minimally adequate treatment for incident depression and anxiety in individuals with inflammatory arthritis (IA).
Methods
We used population-based linked administrative health databases from British Columbia, Canada to evaluate pharmacotherapy and psychological treatments for incident depression and/or anxiety among individuals with IA and without IA (‘IA-free controls’). We defined minimally adequate pharmacotherapy as antidepressant prescriptions filled with ≥ 84 days’ supply and adequate psychological treatment as ≥ 4 counselling/psychotherapy services. Multivariable logistic regression models were used to evaluate the odds of individuals with IA receiving minimally adequate pharmacotherapy and/or psychological treatment compared to IA-free controls.
Results
6,951 (mean age 54.8 ± 18.3 years; 65.5% female) individuals with IA had incident depression and 3,701 (mean age 52.9 ± 16.8 years; 74.3% female) had incident anxiety. Minimally adequate pharmacotherapy and psychological treatment for depression was respectively observed in 50.5% and 19.6% of those with IA, proportions similar to IA-free controls (pharmacotherapy: aOR 1.10, 95% CI 1.00 to 1.21; psychological: aOR 1.07, 95% CI 0.94 to 1.21). Results were similar regarding anxiety treatment. Individuals with IA had a significantly greater likelihood of dispensing ≥ 1 benzodiazepine (anxiety: IA 45.0%, IA-free controls 39.0%, p-value < 0.001) and ≥ 1 tricyclic antidepressant prescription (anxiety: IA 12.8%, IA-free controls 7.8%, p-value < 0.001). Significantly higher average days’ supply of benzodiazepines was observed for IA (anxiety: IA 123.7 days, controls 112.4 days, p-value = 0.003).
Conclusions
A substantial proportion of individuals with IA were not receiving adequate mental health treatment for depression and anxiety, a finding similar for IA-free controls. The undertreatment of mental disorders for people with IA has well-known negative implications for the provision of effective rheumatology care. It remains fundamental to expand publicly funded health care to include mental health services in an effort to address unmet counselling needs.
... Baseline covariates were assessed within 12 months prior to RA index date in the analyses evaluating risk of hip fracture, and within 12 months prior to hip fracture date in the analyses evaluating mortality risk post hip fracture. Covariates included age, sex, social economic status (SES) based on the first three postal code digits and neighborhood income quintile, urban/rural area, the Romano modification of the Charlson comorbidity index (excluding RA from comorbidities) for use with administrative data (cardiovascular disease including coronary artery disease, congestive heart failure, peripheral vascular disease, venous thromboembolism; respiratory conditions; dementia; depression; malignancy; cerebrovascular accident; chronic liver disease/cirrhosis; alcoholism; chronic kidney disease), [30][31][32] and hospital size for postfracture mortality analyses. A single diagnostic code (ICD-9/10) in hospitalization data and physician service claims were used to identify comorbidities (see Supplementary Table 1). ...
Objective
Osteoporosis, a known complication of rheumatoid arthritis (RA), increases the risk of hip fracture, which is associated with high morbidity and mortality. Fracture risk estimates in patients with RA treated with contemporary treatment strategies are lacking. The objectives were (1) estimate age‐specific and sex‐specific incidence rates and compare the risk of hip fractures in RA relative to age‐matched and sex‐matched general population controls, and (2) compare the risk of all‐cause mortality in RA and general population controls after hip fracture.
Methods
A longitudinal study of a population‐based incident cohort of patients with RA diagnosed between 1997 and 2009, followed until 2014, with age‐matched and sex‐matched controls from the general population of British Columbia, using administrative health data. Hip fracture outcomes (International Classification of Diseases, Ninth Edition, Clinical Modification [ICD‐9‐CM] codes 820.0 or 820.2; ICD‐10‐Canada code S72.0 to S72.2) and mortality at predefined intervals after fracture (in hospital, 90 days, 1‐year, 5‐year) were identified. Hip fracture incidence rates for RA and controls, and incidence rate ratios (IRRs), were calculated. Cox proportional hazards models compared hip fracture and mortality risk in RA versus controls; logistic regression compared in‐hospital mortality risk.
Results
Overall, 1,314 hip fractures over 360,521 person‐years were identified in 37,616 individuals with RA and 2,083 over 732,249 person‐years in 75,213 controls, yielding a 28% greater fracture risk in RA (IRR 1.28 [95% confidence interval 1.20–1.37]). Mean age at time of fracture was slightly younger for RA than controls (79.6 ± 10.8 vs 81.6 ± 9.3 years). Postfracture mortality risk at one‐year and five‐years did not differ between RA and general population controls. Results were similar in a sensitivity analysis including only individuals with RA who received disease‐modifying antirheumatic drugs.
Conclusion
People with RA had a greater risk of hip fractures, but no greater risk of mortality post fracture, than the general population. The relative risk of hip fractures observed was not as high as previously reported, likely reflecting better treatment of inflammation and management of osteoporosis and its risk factors.
... Baseline demographic data collected at index included age, gender, geographic region, payer type, and index year. Clinical outcomes assessed were physician specialty type (at office visit on or closest to index date), Dartmouth-Manitoba version of the Charlson Comorbidity Index (DM-CCI) score-a weighted index that uses comorbidity data to predict mortality risk (higher scores = higher risk) [28,29], and preindex prevalence of select conditions of interest, including gout-associated conditions, gout flares, oral ULT, and non-urate lowering goutrelated therapies (e.g., pain medication). Gout flares were identified using a claims-based algorithm based on a published definition using medical and pharmacy claims [30]. ...
Gout is a progressive form of arthritis that causes significant pain and disability. Patients with treatment-refractory (or uncontrolled) gout experience a higher prevalence and severity of comorbidities than those whose gout is controlled. Pegloticase is a recombinant PEGylated uricase indicated for the treatment of gout in patients refractory to conventional therapy. We evaluated the treatment journey of patients with chronic uncontrolled gout before initiation of pegloticase therapy.
Using IQVIA’s PharMetrics® Plus database, we conducted a retrospective observational analysis of adults with ≥ 1 pegloticase claim between April 1, 2011, and August 31, 2020. Demographics were assessed at baseline. Clinical outcomes, health care resource utilization (HCRU), and associated costs were compared over two 12-month periods (months 13–24 and 1–12) prior to the first pegloticase claim (index date).
The study included 408 patients. Prevalence of all gout-associated conditions increased between months 1–12 and 13–24 (P < 0.05 for all). The percentage of patients with tophi increased from 15.4% to 61.5%, the percentage with ≥ 1 flare increased from 49% to 84%, and mean number of flares per patient increased from 1.0 to 2.1 (P < 0.0001 for all). The frequency of all categories of HCRU except emergency department visits also increased (P < 0.0001 for all), as did gout-related healthcare utilization (P£0.005).
Patients with uncontrolled gout experienced an increase in the clinical burden of disease and HCRU in the 2 years before the initiation of pegloticase. Earlier patient identification and initiation of potentially effective therapy may help alleviate these burdens.
... Nonmetropolitan counties were classified as rural (Rural-Urban Continuum and metropolitan counties as urban (Rural-Urban Continuum Codes 1-3) [28]. The National Cancer Institute comorbidity index score was abstracted using claims for comorbid conditions in the year prior to breast cancer diagnosis and classified based on the Klabunde modification for comorbidities [29][30][31][32]. ...
Purpose
Rehabilitation services are recommended by clinical practice guidelines following breast cancer treatment, yet little is known about how utilization may vary by patient-level characteristics which we aimed to study using SEER-Medicare data.
Methods
Data from the Surveillance, Epidemiology, and End Results (SEER)-Medicare linked database was used to identify non-metastatic breast cancer survivors aged ≥ 66 years diagnosed between 2011 and 2016. Rehabilitation services delivered 0–11 months post-diagnosis were identified via outpatient or physician visit claims. Descriptive statistics and associations between patient characteristics and rehabilitation services were calculated using modified Poisson models estimating relative risk (RR) and corresponding 95% confidence intervals (CIs).
Results
Of 55,539 breast cancer survivors, 33% (n = 18,244) had received any type of rehabilitative services. Survivors were a mean age of 75 years (SD 6.7), 88% White, 86% urban-dwelling, and 21% Medicare/Medicaid dually enrolled. In adjusted models, patients aged > 75 vs. ≤ 75 were 6% (RR 0.94, 95% CI 0.92–0.96) less likely to have received rehabilitative services. Survivors in an area with greater educational attainment vs. less educational attainment, White vs. non-White, or living in a rural vs. urban area were 26% (1.26, CI 1.22–1.30), 6% (1.06, CI 1.02–1.11), and 6% (1.06, CI 1.02–1.10) more likely to have received rehabilitative services, respectively.
Conclusion
The largest differences in rehabilitation utilization were observed for survivors of differing educational and treatment statuses.
Implications for Cancer Survivors
Further research is needed on barriers, access, and delivery of rehabilitation services, specifically for breast cancer survivors who are older-aged, non-White, or Medicare/Medicaid dual eligible.
... Covariates included sociodemographic factors (age, sex, neighborhood income quintile [measure of area-level socioeconomic status], and health authority region), comorbidities at any time from enrollment to the index date based on ICD-9 or ICD-10 codes (obesity, hypertension, MI, stroke, transient ischemic attack, congestive heart failure, ischemic heart disease, varicose veins, venous thromboembolism, chronic kidney disease, cancer, infection, and chronic obstructive pulmonary disease), diabetes-related complications (diabetic nephropathy, neuropathy, or retinopathy, and hospitalizations due to diabetes), modified Charlson Comorbidity Index for administrative data [33], medication use (other antidiabetic medications, immunosuppressives [biologics, corticosteroids, other oral and subcutaneous agents], cardiovascular medications [aspirin, statins, fibrates, diuretics, other], proton pump inhibitors, colchicine, urate-lowering therapy, opioids, non-steroidal anti-inflammatory drugs, and anticoagulants), and healthcare utilization (hospitalizations and ER visits) in the 12 months before the index date (Table 1). Diabetes and IMID duration were also included, where date of diagnosis was defined by first ICD-9/10 code (Table 1). ...
Objective
To assess the risk of all-cause mortality and major adverse cardiovascular events (MACE) in patients with immune-mediated inflammatory diseases (IMIDs) and type 2 diabetes newly initiating glucagon-like peptide-1 receptor agonists (GLP-1-RAs) versus dipeptidyl peptidase-4 inhibitors (DPP-4is).
Methods
We performed a population-based cohort study using administrative health data from British Columbia. Patients with an IMID (i.e., rheumatoid arthritis, psoriatic disease, ankylosing spondylitis, inflammatory bowel disease, or a systemic autoimmune rheumatic disease) and type 2 diabetes who newly initiated a GLP-1-RA or DPP-4i between January 1, 2010, and December 31, 2021 were identified using ICD-9/10 codes. The primary outcome was all-cause mortality. Secondary outcomes included MACE and its components (i.e., cardiovascular death, myocardial infarction, and ischemic stroke). Cox proportional hazard regressions were used with propensity score overlap weighting. The analysis was repeated in age- and sex-matched adults without IMIDs.
Results
We identified 10,855 adults with IMIDs and type 2 diabetes who newly initiated a GLP-1-RA or DPP-4i. All-cause mortality rate was lower among initiators of GLP-1-RAs compared to initiators of DPP-4is, with a weighted hazard ratio (HR) of 0.48 (95% confidence interval [CI], 0.31–0.75) and rate difference (RD) of -9.4 (95% CI, -16.0 to -2.7) per 1000 person-years. Rate of MACE was also lower with GLP-1-RA exposure (HR 0.66 [0.50–0.88], RD -10.5 [-20.4 to -0.8]). Effect sizes were similar in adults without IMIDs.
Conclusion
In patients with IMIDs and type 2 diabetes, GLP-1-RA exposure is associated with a lower risk of all-cause mortality and MACE compared to a cardioneutral active comparator.
... Prevalent diabetes refers to diabetes diagnoses made before a cancer diagnosis, whereas incident diabetes diagnoses are those made after a cancer diagnosis. Klabunde's version of the Charlson Comorbidity Index (CCI) was calculated using macro codes provided by the NCI 1 year before the cancer diagnosis (23,24). A binary measure of tobacco use at baseline was obtained from ICD codes (25). ...
Background
The risk of diabetes among Asian, Native Hawaiian, and Pacific Islander (ANHPI) women after breast cancer is unclear. This study estimated the risk of incident type II diabetes in older ANHPI and older non-Hispanic White (NHW) women with breast cancer from the US National Cancer Institute’s Surveillance, Epidemiology, and End Results (SEER) Medicare linked claims.
Methods
A matched cohort of 7122 older ANHPI and 21 365 older NHW women with breast cancer were identified from SEER-Medicare between 2000 and 2017. To assess the risk of incident type II diabetes after breast cancer, hazard ratios (HRs) and 95% confidence intervals (95% CI) were estimated using the Cox proportional-hazards regression model.
Results
During the mean 8 years of follow-up, 9.3% of older women with breast cancer developed incident type II diabetes. In comparison with older NHW women, older ANHPI women without a known history of diabetes had an elevated risk of diabetes after breast cancer, with strong associations observed for Pacific Islander (HR = 3.09, 95% CI = 1.43 to 6.67), Vietnamese (HR = 2.12, 95% CI = 1.33 to 2.36), and Filipino (HR = 2.02, 95% CI = 1.57 to 2.59) women with breast cancer, adjusting for potential confounders. Among ANHPI women with breast cancer, more baseline comorbidities and obesity were risk factors for developing incident type II diabetes.
Conclusion
ANHPI women diagnosed with breast cancer had an elevated risk of type II diabetes compared with older NHW women with breast cancer. Routine monitoring and management of diabetes are warranted in older ANHPI women with breast cancer.
... The original Charlson Comorbidity Index, developed by Charlson et al. [12], is a weighted index used to predict the one-year risk of death for patients with specific comorbid conditions upon hospitalization. Deyo et al. [14] and Romano et al. [15] adapted the index to ICD-9-CM diagnosis and procedure codes and CPT-4 codes, respectively, enabling its calculation using administrative data. In our case, the weights for the comorbidity score were determined by rounding off the coefficients obtained from a regression model that utilized "out- for females and males, respectively. ...
Accurately estimating the morbidity of individuals with specific conditions is crucial for healthcare systems to enhance healthcare delivery. In this study, we introduce DKABio-clusters and elucidate their characteristics, demonstrating their significant implications for healthcare management. The key applications of DKABio-clusters include the development of a health score (DKABio-HS) and risk prediction. The health score enables the calculation of a comprehensive "disease-related" score that summarizes an individual's health status into a single value. Our findings reveal that the health score exhibits high consistency, as indicated by minimal differences in health scores between the training and validation samples (mean absolute percentage errors within 0 to 10 years were less than 0.1%, with all mean absolute percentage errors ranging between 1.2-1.6%). A higher health score signifies better health or lower disease risk, with the health score decreasing with age or the presence of multiple diseases. By utilizing the health score, we can establish a classification rule known as the "disease map," enabling quantitative differentiation of individuals at various non-diseased or diseased levels. Consequently, non-disease individuals can be categorized as healthy or sub-healthy, allowing tailored health management approaches for preventive measures. Our results indicate that individuals classified as sub-healthy exhibit significantly higher disease risks compared to healthy individuals (Female (male) 5-year risks of developing at least one disease are 29% vs. 15% (29% vs. 16.5%)). Additionally, by employing a selected set of health variables, we can estimate the distribution of DKABio-clusters and simultaneously predict the 10-year risks of 15 diseases/conditions. We validate the predictive power of our model by comparing the predicted risks with true risks derived from the training and validation samples. In the majority of cases, the differences between the true and predicted risks were not statistically significant. All study findings were based on data from the National Health Insurance Research Database released by the National Health Research Institute, Taiwan, and the Mei Jau Health Management Institution database, collected between 2000 and 2016 in Taiwan.
... COPD severity was graded based on the American Thoracic Society categories from 2005. The Romano adaptation of the Charlson comorbidity index was calculated using International Classification of Diseases, Ninth Revision (ICD-9) codes to develop a composite comorbidity score (22). ...
Background
Locally advanced non-small cell lung cancer (LA-NSCLC) treated with the programmed death-ligand 1 inhibitor durvalumab has been associated with significant rates of pneumonitis, which has led to higher rates of discontinuation of therapy in real-world populations. Thus far there has been no consensus in the literature on the impact of pneumonitis on survival.
Methods
This is a retrospective cohort study of veterans receiving durvalumab between 12/5/2017 and 4/15/2020. Participants were identified using VINCI data services. Patients were followed through 9/14/2021. Development of clinical pneumonitis was assessed through review of documentation and graded using CTCAE 4.0 criteria. Univariate logistic regression analysis evaluated for associations between body mass index (BMI), age, race, co-morbidity index, chemotherapy regimen, chronic obstructive pulmonary disease (COPD) severity, and development of clinical pneumonitis. Progression-free survival (PFS) and overall survival (OS) were evaluated using Kaplan-Meier methods. Cox proportional hazards models were utilized to evaluate the association between risk of death at 1 and 2 years and candidate predictor variables.
Results
A total of 284 patients were included in this study. Sixty-one patients developed clinically significant pneumonitis, 7 patients developed grade 5 pneumonitis (death from pneumonitis). The median OS in patients that developed pneumonitis was 27.8 vs. 36.9 months in patients that did not develop pneumonitis (P=0.22). BMI was found to be a clinical predictor of pneumonitis (P=0.04). COPD severity, race, age at durvalumab start date, chemotherapy regimen, and Romano comorbidity index were not significant predictors of pneumonitis. Cox proportional hazards analysis failed to demonstrate an association between the development of pneumonitis and risk of death in this population.
Conclusions
The incidence of clinically significant pneumonitis is higher than noted in the PACIFIC trial in this cohort, however this high rate of pneumonitis does not have an impact on OS or PFS. Obesity was found to be a significant predictor of pneumonitis in this patient population.
... To control for potential confounding, large-scale propensity score matching (PSM) was employed between the two groups. Covariates used in the matching process included age, sex, prior conditions, drugs observed during the long-term (within ~ 365 days) and short-term (within ~ 30 days) before study drug exposure, and the Romano Adaptation of the Charlson comorbidity index 25 . ...
This observational study explored the association between proton pump inhibitor (PPI) and histamine-2 receptor antagonist (H2RA) use and the risk of chronic kidney disease (CKD). Using the National Health Insurance Service-National Sample Cohort (NHIS-NSC) and six-hospital electronic health record (EHR) databases, CKD incidence was analyzed among PPI and H2RA users. Propensity score matching was used to balance baseline characteristics, with 1,869 subjects each in the PPI and H2RA groups from the NHIS-NSC, and 5,967 in EHR databases. CKD incidence was similar for both groups (5.72/1000 person-years vs. 7.57/1000 person-years; HR = 0.68; 95% CI, 0.35–1.30). A meta-analysis of the EHR databases showed no significant increased CKD risk associated with PPI use (HR = 1.03, 95% CI: 0.87–1.23). These results suggest PPI use may not increase CKD risk compared to H2RA use, but the potential role of PPI-induced CKD needs further research. Clinicians should consider this when prescribing long-term PPI therapy.
... [19] We identified the diagnosis and procedure with the use of the International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM), and the International Classification of Diseases, Ninth Revision, Procedure Coding System. [20] ...
Background:
The recurrence rate of shoulder dislocation is high, and the male sex and young age have been recognized as risk factors. A personalized model for predicting the probability of shoulder redislocation would be useful in clinical practice.
Objectives:
We aimed to establish a prediction model for estimating the redislocation probability within 5 years after nonsurgical management for the first-time episode in young patients.
Materials and Methods:
With the use of a nationwide administrative database, patients aged 13–50 years with a diagnosis of primary shoulder dislocation receiving closed reduction from January 1, 2000, to December 31, 2013, were extracted. We apportioned the data into training and validation sets with a 2:1 split. The putative prognostic predictors, including age, sex, diabetes mellitus, hypertension, hyperlipidemia, and Charlson Comorbidity Score, were identified and analyzed in a nomogram. Penalized Cox regression with the adaptive elastic-net variable selection method was utilized to construct a nomogram.
Results:
A total of 461 patients were included in this study. After regression analysis, age and sex were incorporated into the nomogram. After the model training, a nomogram for prediction was developed; the AUC of this model ranged from 0.704 to 0.751. The calibration plots for the endpoints showed optimal agreement between the nomogram’s prediction and the actual observation.
Conclusions:
The present study proposed a nomogram that can be used to predict the probability of shoulder redislocation within 5 years after nonsurgical treatment for the first shoulder dislocation.
... Assessment of covariates. Covariates included age, sex, rheumatic disease diagnosis, socioeconomic status (neighborhood income quintile based on census data), number of previous COVID-19 PCR tests as a proxy for situations that increase risk of infection, rural address, Romano modification of the Charlson comorbidity index (which includes diabetes) (28), health authority in BC, specific comorbidities, including hypertension, and specific outpatient pharmacy-dispensed medications, including medications for diabetes, health care use (outpatient visits and hospitalization), and month/year of initial positive SARS-CoV-2 test (all covariates listed in Table 1 with the exception of specific comorbidities and outpatient pharmacy-dispensed medications, which are in Supplementary Table 1 and month/year of initial SARS-CoV-2 test, which is in Supplementary Table 2). Covariates were examined in the year before positive SARS-CoV-2 test. ...
Objective:
We estimated the association between immunosuppressive and immunomodulatory agent (IIA) exposure and severe COVID-19 outcomes in a population-based cohort study.
Methods:
Participants were 18 years or older, tested positive for SARS-CoV-2 between February 6, 2020, and August 15, 2021, and were from administrative health data for the entire province of British Columbia, Canada. IIA use within 3 months prior to positive SARS-CoV-2 test included conventional disease-modifying antirheumatic drugs (antimalarials, methotrexate, leflunomide, sulfasalazine, individually), immunosuppressants (azathioprine, mycophenolate mofetil/mycophenolate sodium [MMF], cyclophosphamide, cyclosporine, individually and collectively), tumor necrosis factor inhibitor (TNFi) biologics (adalimumab, certolizumab, etanercept, golimumab, infliximab, collectively), non-TNFi biologics or targeted synthetic disease-modifying antirheumatic drugs (tsDMARDs) (rituximab separately from abatacept, anakinra, secukinumab, tocilizumab, tofacitinib and ustekinumab collectively), and glucocorticoids. Severe COVID-19 outcomes were hospitalizations for COVID-19, ICU admissions, and deaths within 60 days of a positive test. Exposure score-overlap weighting was used to balance baseline characteristics of participants with IIA use compared with nonuse of that IIA. Logistic regression measured the association between IIA use and severe COVID-19 outcomes.
Results:
From 147,301 participants, we identified 515 antimalarial, 573 methotrexate, 72 leflunomide, 180 sulfasalazine, 468 immunosuppressant, 378 TNFi biologic, 49 rituximab, 144 other non-TNFi biologic or tsDMARD, and 1348 glucocorticoid prescriptions. Risk of hospitalizations for COVID-19 was significantly greater for MMF (odds ratio [95% CI]): 2.82 [1.81-4.40], all immunosuppressants: 2.08 [1.51-2.87], and glucocorticoids: 1.63 [1.36-1.96], relative to nonuse. Similar outcomes were seen for ICU admission and MMF: 2.52 [1.34-4.74], immunosuppressants: 2.88 [1.73-4.78], and glucocorticoids: 1.86 [1.37-2.54]. Only glucocorticoids use was associated with a significant increase in 60-day mortality: 1.58 [1.21-2.06]. No other IIAs displayed statistically significant associations with severe COVID-19 outcomes.
Conclusion:
Current use of MMF and glucocorticoids were associated with an increased risk of severe COVID-19 outcomes compared with nonuse. These results emphasize the variety of circumstances of patients taking IIAs.
... To control for potential confounding, large-scale propensity score matching (PSM) was employed between the two groups. Covariates used in the matching process included age, sex, prior conditions, drugs observed during the long-term (within ~ 365 days) and short-term (withiñ 30 days) before study drug exposure, and the Romano Adaptation of the Charlson comorbidity index 23 . Propensity scores were estimated using logistic regression models with an L1 penalty hyperparameter, selected through 10-fold cross-validation 24 . ...
This observational study explored the association between proton pump inhibitor (PPI) and histamine-2 receptor antagonist (H2RA) use and the risk of chronic kidney disease (CKD). Using the National Health Insurance Service-National Sample Cohort (NHIS-NSC) and six-hospital electronic health record (EHR) databases, CKD incidence was analyzed among PPI and H2RA users. Propensity score matching was used to balance baseline characteristics, with 1,858 subjects each in the PPI and H2RA groups from the NHIS-NSC, and 5,772 in EHR databases. CKD incidence was similar for both groups (7.39/1000 person-years vs. 7.03/1000 person-years; HR, 1.00; 95% CI, 0.56-1.79). A meta-analysis of the EHR databases showed no significant increased CKD risk associated with PPI use (HR = 1.01, 95% CI: 0.85-1.20). These results suggest PPI use may not increase CKD risk compared to H2RA use, but the potential role of PPI-induced CKD needs further research. Clinicians should consider this when prescribing long-term PPI therapy.
... Each outcome model accounted for confounders including age, sex, Charlson Comorbidity Index (CCI), number of days of available data up to first SCS prescription, and body mass index (BMI). The original 19-category CCI was modified to 17 categories [31] and the list of specific ICD diagnosis codes used to identify different categories of comorbidity were modified [32] and updated from ICD-9 to ICD-10 coding [33,34]. ...
IntroductionThis analysis was conducted to assess the incidence of adverse clinical outcomes, healthcare resource use (HCRU), and the costs associated with systemic corticosteroid (SCS) use in adults with systemic lupus erythematosus (SLE) in the UK.Methods
We identified incident SLE cases using the Clinical Practice Research Datalink GOLD, Hospital Episode Statistics-linked healthcare, and Office for National Statistics mortality databases from January 1, 2005, to June 30, 2019. Adverse clinical outcomes, HCRU, and costs were captured for patients with and without prescribed SCS.ResultsOf 715 patients, 301 (42%) had initiated SCS use (mean [standard deviation (SD)] 3.2 [6.0] mg/day) and 414 (58%) had no recorded SCS use post-SLE diagnosis. Cumulative incidence of any adverse clinical outcome over 10-year follow-up was 50% (SCS group) and 22% (non-SCS group), with osteoporosis diagnosis/fracture most frequently reported. SCS exposure in the past 90 days was associated with an adjusted hazard ratio of 2.41 (95% confidence interval 1.77–3.26) for any adverse clinical outcome, with increased hazard for osteoporosis diagnosis/fracture (5.26, 3.61–7.65) and myocardial infarction (4.52, 1.16–17.71). Compared to low-dose SCS (< 7.5 mg/day), patients on high-dose SCS (≥ 7.5 mg/day) had increased hazard for myocardial infarction (14.93, 2.71–82.31), heart failure (9.32, 2.45–35.43), osteoporosis diagnosis/fracture (5.14, 2.82–9.37), and type 2 diabetes (4.02 1.13–14.27). Each additional year of SCS use was associated with increased hazard for any adverse clinical outcome (1.15, 1.05–1.27). HCRU and costs were greater for SCS users than non-SCS users.Conclusions
Among patients with SLE, there is a higher burden of adverse clinical outcomes and greater HCRU in SCS versus non-SCS users.
... 38 The Charlson index has been adapted for its use in administrative claims databases. [39][40][41] This study uses the Deyo's modification of the Charlson index as a measure to assess the overall severity of the illness in the study population. ...
Objective: Adults with developmental disabilities have higher prevalence of chronic disease conditions such as diabetes, obesity, high blood pressure, arthritis, CVD and chronic pain. They also have poor healthcare utilization and are screened less for chronic disease conditions. Very few studies have looked at the diabetes related health outcomes in developmentally disabled adults, where distinct racial differences in medication use behaviors have been observed. The objective of this study was to examine the association of race with healthcare utilization and costs in pharmacologically treated Medicaid enrollees with developmental disabilities (DD) and type 2 diabetes. Methods: This was a retrospective cohort study that identified adults with DD and type 2 diabetes from the MarketScan® Multi-State Medicaid Database. Enrollees aged 18-64 years who received new medications for type 2 diabetes from January 1, 2004 and December 31, 2006 were included. An index diagnosis date was assigned to each patient and adults with a continuous enrollment for at least 12 months were included. Probabilities of type 2 diabetes related healthcare utilization (inpatient, outpatient and emergency department visits) in adults with DD were computed using multivariate logistic regression models. Multivariate negative binomial regression was used to measure the rate of change in type 2 diabetes related healthcare utilization in patients with DD. Multivariate linear regression with log-transformation was used to determine type 2 diabetes related healthcare costs in Medicaid enrollees with DD. Results: This study had a sample size of 1529 patients. After controlling for all the covariates, compared to Caucasians with DD, African Americans with DD were more likely to have type 2 diabetes related inpatient (OR=1.71; 95% CI, 1.02-2.85) and emergency department visits (OR, 1.67; 95% CI, 1.02-2.73). African Americans with DD and type 2 diabetes had significantly higher healthcare costs compared to Caucasians with DD and type 2 diabetes. Conclusion: Racial disparities exist in healthcare utilization in Medicaid patients with DD and type 2 diabetes. African Americans were more likely to have inpatient and ER visits respectively. Also, African Americans were more likely to have higher type 2 related healthcare costs compared to Caucasians.
... The original version of the CCI was based on 19 items corresponding to different clinical comorbidities [4]. Subsequently, different versions of the CCI have been developed based on different sources of data, including the age-adjusted CCI, ICD-9 code based CCI [6][7][8] and ICD-10 code based CCI [9,10]. ...
Background
Cardiac arrest is currently one of the leading causes of mortality in clinical practice, and the Charlson Comorbidity Index (CCI) is widely utilized to assess the severity of comorbidities. We aimed to evaluate the relationship between the age-adjusted CCI score and in-hospital mortality in intensive care unit (ICU) patients with the diagnosis of cardiac arrest, which is important but less explored previously.
Methods
This was a retrospective study including patients aged over 18 years from the MIMIC-IV database. We calculated the age-adjusted CCI using age information and ICD codes. The univariate analysis for varied predictors’ differences between the survival and the non-survival groups was performed. In addition, a multiple factor analysis was conducted based on logistic regression analysis with the primary result set as hospitalization death. An additional multivariate regression analysis was conducted to estimate the influence of hospital and ICU stay.
Results
A total of 1772 patients were included in our study, with median age of 66, among which 705 (39.8%) were female. Amongst these patients, 963 (54.3%) died during the hospitalization period. Patients with higher age-adjusted CCI scores had a higher likelihood of dying during hospitalization ( P < 0.001; OR: 1.109; 95% CI: 1.068–1.151). With the age-adjusted CCI incorporated into the predictive model, the area under the receiver operating characteristic curve was 0.794 (CI: 0.773–0.814), showing that the prediction model is effective. Additionally, patients with higher age-adjusted CCI scores stayed longer in the hospital ( P = 0.026, 95% CI: 0.056–0.896), but there was no significant difference between patients with varied age-adjusted CCI scores on the days of ICU stay.
Conclusion
The age-adjusted CCI is a valid indicator to predict death in ICU patients with cardiac arrest, which can offer enlightenment for both theory literatures and clinical practice.
... Predictors and confounders. We selected known SSI risk factors for inclusion in the data analysis such as age-adjusted Charlson Comorbidity Index scores, sex, length of stay, surgery procedure category, MSA, and Medicare status [34][35][36][37] . In January 2012, a new mandate required hospitals to report their SSIs to the NHSN, which the Centers for Medicare and Medicaid Services used to qualify hospitals for their annual payment update 38 . ...
Surgical site infections (SSI) are one of the most common and costly hospital-acquired infections in the United States. Meteorological variables such as temperature, humidity, and precipitation may represent a neglected group of risk factors for SSI. Using a national private insurance database, we collected admission and follow-up records for National Healthcare Safety Network-monitored surgical procedures and associated climate conditions from 2007 to 2014. We found that every 10 cm increase of maximum daily precipitation resulted in a 1.09 odds increase in SSI after discharge, while every g/kg unit increase in specific humidity resulted in a 1.03 odds increase in SSI risk after discharge. We identified the Southeast region of the United States at highest risk of climate change-related SSI, with an estimated 3% increase in SSI by 2060 under high emission assumptions. Our results describe the effect of climate on SSI and the potential burden of climate-change related SSI in the United States.
... A total of 16 established indices or measures of comorbidity were used across the papers reviewed, with 38 studies using or including a measure of comorbidity developed specifically for the published study. The number of papers using each index or measure is presented in Fig. 2, with Table I outlining basic information about the indices identified during data extraction where available (28)(29)(30)(31)(32)(33)(34)(35)(36)(37)(38)(39)(40)(41). Of these, the CCI was the most frequently utilised, with 65 (41.1%) ...
Epidemiological and retrospective clinical studies on cancer outcomes frequently adjust for patients' comorbid conditions. Despite the existence of multiple comorbidity indices, the Charlson comorbidity index (CCI) is the most frequently applied. Indices are developed in specific settings and the extent of alignment between the development setting and subsequent study is unclear. The present study provides a contemporaneous snapshot of comorbidity indices used in retrospective observational cancer studies and the extent to which cancer type(s), data source(s) and outcome(s) matched the studies in which the indices were developed. A systematic literature search in PubMed identified retrospective, observational studies on outcomes in patients with cancer published between March 2015 and March 2020. Information including the cancer type, data source and outcome were extracted and compared to those used in the validation study of the comorbidity index used. Of 158 papers reviewed, 79 used the CCI, either alone or in combination with other indices. The cancer type matched to that used in the validation study of the comorbidity index in 16 of the 115 studies using an established index, whilst the data source matched in 27 studies and outcome in only two. Justification was rarely provided for index choice (15 of the 115 studies). It may be concluded that, while the CCI remains the dominant comorbidity index, it may not always align to key elements of the study design in terms of cancer type, data source and outcome. A range of indices exists and identification of the most appropriate measure has the potential to improve adjustment for comorbidity. The present study provided information about the indices used in included studies and encourages future studies to consider which comorbidity index offers the best alignment with the study population, data source and question addressed.
... The covariates that we assessed in the year prior to the index date were age, sex, neighbourhood income quintile, residence (urban/rural) (Additional file 1), comorbidities as measured using the Charlson-Romano Comorbidity Index [16], and health care utilization characterized by the number of outpatient, inpatient, and rheumatology visits. ...
Background
Gout may be associated with an increased incidence of mental health disorders, however, published findings have been limited and inconsistent. Therefore, our objective was to conduct a population-based cohort study to evaluate the incidence of depression and anxiety after gout diagnosis.
Methods
We used linked population-based administrative health data in British Columbia, Canada that includes information on demographics, outpatient visits, and inpatient visits from the period of January 1, 1990 to March 31, 2018. We assessed depression and anxiety using validated International Classification of Diseases, 9th and 10th Revision coding algorithms. We applied multivariable Cox proportional hazard models to evaluate incident depression and anxiety among patients with gout in comparison to non-gout controls, adjusting for age, sex, neighbourhood income quintile, residence, comorbidities, and health care utilization.
Results
We included 157,426 incident cases of gout (60.2% male; mean age 57.1 years) and 157,426 non-gout controls (60.2% male; mean age 56.9 years). The incidence rate of depression among individuals with gout and non-gout controls was 12.9 (95% confidence interval [CI] 12.7–13.2) and 11.1 (95% CI 10.9–11.4) per 1000 person-years, respectively. The incidence rate of anxiety for those with gout was 5.4 (95% CI 5.3–5.5) per 1000 person-years and for non-gout controls was 4.6 (95% CI 4.4–4.7) per 1000 person-years. Individuals with gout had an increased onset of depression (adjusted hazard ratio [aHR], 1.08; 95% CI 1.05–1.11) and anxiety (aHR, 1.10; 95% CI 1.05–1.14) compared to non-gout controls.
Conclusion
Our population-based study shows an increased incidence of depression and anxiety following gout diagnosis in comparison to non-gout controls. Findings suggest the importance of considering psychiatric impacts in addition to the physical impacts of gout.
... To account for morbidity and mortality risk, we adjusted for the Romano-Charlson Comorbidity Index (21). A higher comorbidity index indicates worse health. ...
Objective:
Electronic health record (EHR) databases are a powerful resource to investigate clinical trajectories of osteoarthritis (OA). There are no existing EHR tools to evaluate risk for knee arthroplasty (KA). We developed an OA severity index (OASI) using EHR data and demonstrate the index's association with time to KA.
Methods:
This retrospective cohort study used 2010-2018 nationally distributed Optum EHR data. Eligible patients were 45 to 80 years old with a new diagnosis of knee OA in 2011-2012 and no prior KA. The OASI was a sum of first instance of x-ray imaging, advanced imaging, intra-articular injection, nonsteroidal anti-inflammatory drugs, and opioids. Principal components analysis index (PCI) score was also explored. Extended Cox proportional hazard models assessed time-dependent OASI and time to KA.
Results:
Among 16,675 eligible patients, 12.7% underwent KA. Median follow-up time was 72 months. Adjusted OASI models showed each additional event almost doubled the risk for KA (adjusted hazard ratio = 1.80, 95% confidence interval: 1.75-1.86). Similar results were observed for PCI.
Conclusion:
The sum OASI performs well identifying patients who would undergo KA and offers simplicity versus the PCI. Although replication in other cohorts is recommended, the OASI appears to be a novel and valid means to measure clinical OA severity in research studies using large EHR-based cohorts.
... 19,20 To account for morbidity and mortality risk, we adjusted for the Romano -Charlson Comorbidity Index. 21 A higher comorbidity index indicates worse health. New opioid prescriptions included long-acting and short-acting formulations for the medications shown in Supplemental Table 1. ...
Objective
Black patients are less likely than White patients to receive physical therapy for musculoskeletal pain conditions. Current evidence, however, is limited to self-reported conditions and health services use. The purpose of this study was to use a large electronic health record data base to determine whether a race disparity existed in use of physical therapy within 90 days of a new musculoskeletal diagnosis.
Methods
Eligible patients (n = 52,384) were sampled from a Optum deidentified electronic health record (EHR) database of 5 million adults distributed throughout the United States. In this database, patients were designated as “Black” and “White.” Patients were eligible if they had a new diagnosis for musculoskeletal neck, shoulder, back, or knee pain between 1/1/2012 and 12/31/2017. Logistic regression and Cox proportional hazard models were computed before and after adjusting for covariates to estimate the association between race and receipt of physical therapy services within 90 days of musculoskeletal pain diagnoses.
Results
Patients were on average 47.5 (SD = 14.9) years of age, 12.8% were Black, 87.2% were White, and 52.7% were female. Ten percent of Black patients and 15.5% of White patients received physical therapy services within 90 days of musculoskeletal pain diagnoses. After adjusting for covariates, White patients were 57% more likely (odds ratio [OR] = 1.57; 95% CI = 1.44–1.71) to receive physical therapy as compared with Black patients and had significantly shorter time to physical therapy as compared with Black patients (hazard ratio [HR] = 1.53; 95% CI = 1.42–1.66).
Conclusions
In a nationally distributed cohort, Black patients were less likely than White patients to utilize physical therapy and had a longer time to utilization of physical therapy for musculoskeletal pain.
... 3 Although widely adopted, the Charlson Comorbidity Index (CCI) was criticized based on its breadth and applicability to administrative health data. 4,5 The Elixhauser Comorbidity Index (ECI) was developed in 1998 to contain more disease categories (30) and summarize more conditions than the CCI. 6 The ECI was updated in 2012 to include 31 comorbidity disease categories, 7 and it has been shown to be more versatile and statistically rigorous than the CCI. ...
Objective
The goal of this study was to investigate the relationship between comorbidities and ethnic health disparities in a diverse, cosmopolitan population.
Materials and Methods
We used the UK Biobank (UKB), a large progressive cohort study of the UK population. Study participants self-identified with 1 of 5 ethnic groups and participant comorbidities were characterized using the 31 disease categories captured by the Elixhauser Comorbidity Index. Ethnic disparities in comorbidities were quantified as the extent to which disease prevalence within categories varies across ethnic groups and the extent to which pairs of comorbidities co-occur within ethnic groups. Disease-risk factor comorbidity pairs were identified where one comorbidity is known to be a risk factor for a co-occurring comorbidity.
Results
The Asian ethnic group shows the greatest average number of comorbidities, followed by the Black and then White groups. The Chinese group shows the lowest average number of comorbidities. Comorbidity prevalence varies significantly among the ethnic groups for almost all disease categories, with diabetes and hypertension showing the largest differences across groups. Diabetes and hypertension both show ethnic-specific comorbidities that may contribute to the observed disease prevalence disparities.
Discussion
These results underscore the extent to which comorbidities vary among ethnic groups and reveal group-specific disease comorbidities that may underlie ethnic health disparities.
Conclusion
The study of comorbidity distributions across ethnic groups can be used to inform targeted group-specific interventions to reduce ethnic health disparities.
... Patients who received early PT were significantly more likely to be female and were slightly younger than those who did not receive early PT (Table 1). We calculated Charlson Comorbidity Index (CCI) values for each patient [17]. Among patients receiving early PT, a significantly smaller proportion had a CCI value ≥1 (1.8%) compared with those not receiving early PT (2.6%) (P < 0.001). ...
Abstract Background Early initiation of physical therapy (PT) has been associated with lower healthcare costs and utilization; however, these studies have been limited to single institutions or healthcare systems. Our goal was to assess healthcare utilization and spending among patients who present for the first time with low back pain (LBP), according to whether they received early physical therapy (PT), using a large, nationwide sample; and geographic variation in rates of early PT and 30-day LBP-related spending. Methods Using the Truven MarketScan database, we identified nearly 980,000 US adults ages 18–64 years who initially presented with acute LBP from 2010 through 2014 and did not have nonmusculoskeletal causes of LBP. Approximately 110,000 patients (11%) received early PT (≤2 weeks after presentation). We compared healthcare utilization and spending at 30 days and 1 year after presentation between patients who received early PT and those who did not. Alpha = 0.05. Results At 30 days, early PT was associated with lower odds of chiropractor visits (odds ratio [OR] = 0.41, 95% confidence interval [CI] = 0.40–0.42), pain specialist visits (OR = 0.49, 95% CI = 0.47–0.51), emergency department visits (OR = 0.51, 95% CI = 0.49–0.54), advanced imaging (OR = 0.57, 95% CI = 0.56–0.58), orthopaedist visits (OR = 0.67, 95% CI = 0.66–0.69), and epidural steroid injections (OR = 0.68, 95% CI = 0.65–0.70). At 1 year, early PT was associated with less healthcare utilization. At 30 days, patients with early PT had lower mean LBP-related spending (1500) compared with those without early PT (2560) (P
... For drug groups, we separately considered data from the 30 days prior to index and data from the 365 days prior to index. Finally, three established risk scores (CHA2DS2-VASc, Diabetes Complications Severity Index (DCSI), Charlson Comorbidity Index (CCI) (Romano adaptation)) were calculated using all data prior to index [23][24][25][26]. No clinical measurements or sporadically recorded variables were included as candidate predictors to maximize transportability (i.e., ability to apply across databases) of the developed prediction models. ...
Background
Identification of rheumatoid arthritis (RA) patients at high risk of adverse health outcomes remains a major challenge. We aimed to develop and validate prediction models for a variety of adverse health outcomes in RA patients initiating first-line methotrexate (MTX) monotherapy.
Methods
Data from 15 claims and electronic health record databases across 9 countries were used. Models were developed and internally validated on Optum® De-identified Clinformatics® Data Mart Database using L1-regularized logistic regression to estimate the risk of adverse health outcomes within 3 months (leukopenia, pancytopenia, infection), 2 years (myocardial infarction (MI) and stroke), and 5 years (cancers [colorectal, breast, uterine] after treatment initiation. Candidate predictors included demographic variables and past medical history. Models were externally validated on all other databases. Performance was assessed using the area under the receiver operator characteristic curve (AUC) and calibration plots.
Findings
Models were developed and internally validated on 21,547 RA patients and externally validated on 131,928 RA patients. Models for serious infection (AUC: internal 0.74, external ranging from 0.62 to 0.83), MI (AUC: internal 0.76, external ranging from 0.56 to 0.82), and stroke (AUC: internal 0.77, external ranging from 0.63 to 0.95), showed good discrimination and adequate calibration. Models for the other outcomes showed modest internal discrimination (AUC < 0.65) and were not externally validated.
Interpretation
We developed and validated prediction models for a variety of adverse health outcomes in RA patients initiating first-line MTX monotherapy. Final models for serious infection, MI, and stroke demonstrated good performance across multiple databases and can be studied for clinical use.
Funding
This activity under the European Health Data & Evidence Network (EHDEN) has received funding from the Innovative Medicines Initiative 2 Joint Undertaking under grant agreement No 806968. This Joint Undertaking receives support from the European Union's Horizon 2020 research and innovation programme and EFPIA.
... We used data from the Japan Medical Data Center (JMDC), a payerbased database that has collected claims from more than 250 payers and covers workers and their dependents and from two US data- 7 Balance between early and mid-adopters and between early and later adopters was assessed using standardized mean differences (SMDs). An SMD >0.1 was used as an ad hoc criterion for a substantial difference. ...
Early Post‐Marketing Phase Vigilance (EPPV) is a unique system that encourages reporting of serious adverse reactions for medications newly introduced to Japan. When a once‐monthly paliperidone palmitate formulation (PP1M) was introduced in Japan in 2013, EPPV detected a signal of increased mortality, but this signal was not subsequently confirmed. To clarify whether that signal reflected increased adverse event reporting or an atypically high baseline mortality risk among early adopters of PP1M, we evaluated the baseline risk characteristics of early, mid, and later adopters of PP1M in a Japanese database and did a similar evaluation of PP1M and the three‐monthly formulation (PP3M) in two US databases. In Japan, early adopters compared with later adopters were older (mean 39.16 vs 33.70 years) but had a lower proportion of male patients (32.0% vs 44.44%), and a lower mean number of antipsychotic medications (distinct active medical substances) other than paliperidone (2.62 vs 2.85). In the United States, the baseline characteristics of early adopters of PP1M and PP3M did not suggest higher mortality risk than later adopters. These results offer no convincing evidence that the unconfirmed early signal of increased mortality with PP1M was due to increased baseline mortality risk among early adopters.
... Student's t tests were applied for continuous variables and Chi-Square tests for dichotomous variables. We used a separate CCI is a widely accepted comorbidity summary measure, which is included in mortality-predicting models in order to account for variability in the known burden of patient comorbidities between periods [10,11]. Analyses were performed using IBM SPSS Statistics v25. ...
Background:
A substantial drop in emergency department (ED) visit volume was previously demonstrated at the onset of the COVID-19 pandemic.
Objective:
To examine changes in the number of non-COVID adult ED visits and their associated 30-day mortality during 14 months of the pandemic in Israel.
Methods:
This is a retrospective cohort study including 1,285,270 adult ED visits between 1st March, 2018 and 30th April, 2021 to the internal and surgical EDS in eight general hospitals of the largest healthcare organization in Israel. The 14 months of the pandemic period (March 2020-April 2021) were divided into seven periods according to dates of the three lockdowns. Exposure to each of these periods was compared to the parallel period during the two previous years. March 2020-April 2021 was compared to the parallel periods in 2018 and 2019.
Results:
During the pandemic period, the largest decline in ED visits (44.6% and 50.9% for internal and surgical EDs, respectively) and the highest excess 30-day mortality following an ED visit (internal EDs Adjusted OR (ORadj), 1.49; 95% CI, 1.34-1.66 and surgical EDs: ORadj 1.50; CI, 1.16-1.94) were 95%, observed during the first lockdown. Both gradually levelled-off subsequently until near-normalization was reached in March-April 2021 for both parameters.
Conclusions:
A substantial decline in non-COVID ED visits and excess mortality at the beginning of the pandemic, are probably the results of social distancing restrictions alongside patients' fear of exposure to COVID-19, which gradually moderated thereafter, until near normalization was reached after 14 months. Gradual return to pre-pandemic ED utilization patterns were noticed as the population and the healthcare system acclimatize to life alongside COVID.
... Age, gender, and treatment year were derived from source datasets; cancer stage, grade, and tumor location were available in SEER/TCR. Claims-derived covariates included: modified Charlson comorbidity score based on diagnosis claims and performance status indicators based on durable medical equipment claims [7][8][9][10][11][12]13] (eTable 1). ...
Background
We assessed radiation treatment (RT) use and complications for unresectable pancreatic cancer in the US, comparing conventionally fractionated (CFRT) and stereotactic body radiation treatment (SBRT) to inform real-world expected outcomes and practice.
Material and Methods
We analyzed 5,624 patients with non-metastatic, unresectable pancreatic cancer (2,522 older patients age>65, diagnosed 2006-2013 in Medicare linked data; and 3,102 younger patients age<65, diagnosed 2006-2016 in MarketScan data), comparing CFRT vs. SBRT vs. chemotherapy alone. Cochran-Armitage tested temporal trends. Fisher’s Exact Test and proportional hazards models compared gastrointestinal (GI) complications. Healthcare payments (Consumer Price Index adjusted to 2015) through 12 months were compared using generalized linear regression models with log link and gamma distribution.
Results
RT use declined from 55% to 45% of older patients (2006-2013) and 52% to 47% of younger patients (2006-2016) (Ptrend<0.001 both). Among RT patients, SBRT use increased to 10% of older patients and 12% of younger patients in the most recent years (Ptrend=0.04 and <0.001 respectively). Addition of RT was associated with more frequent GI bleeds, strictures, and fistulas (Δ= +3% to 9% excess events, all P≤0.05). Temporal patterns suggested decreasing complications over time (Ptrend=0.05 and 0.05 for older and younger patients). Among younger patients, there was no difference in GI complications for SBRT vs. CFRT (P>0.05, all comparisons). Among older patients, increased complications were seen for SBRT in 1-4 fractions vs. CFRT (P<0.05), but not SBRT in 5 fractions (P=0.72). Healthcare payments were greatest for SBRT when compared with CFRT or chemotherapy under US Medicare (P<0.001) and employer-based insurance (P<0.001).
Conclusion
Real-world treatment has shifted toward more selectivity for RT in unresectable pancreatic cancer. However, SBRT uptake and improving trends in complications profiles represent opportunities to optimize current use and benefit. Findings are applicable to inform future comparative and cost effectiveness models of RT for this disease.
... Data were retrieved using a standardized spreadsheet provided to the examiners in each participating center. The following demographics and clinical variables were retrieved: age, sex, body mass index (BMI), American Society of Anesthesiologists (ASA) class, comorbidity, Charlson Comorbidity Index [25,26], tumor location, and tumor stage (AJCC). Histopathological variables included tumor size, pT stage, pN stage, R0 resection, number of lymph nodes retrieved, perineural and perivascular invasion, and tumor grade. ...
Purpose:
The optimal approach for minimally invasive (MIS) right colectomy remains under debate. This study aimed to describe surgical trends in the treatment of nonmetastatic right colon cancer and to identify predictors of short-term surgical outcomes.
Methods:
A retrospective multicenter cohort study of Minimally-invasivE surgery for oncologic Right ColectomY (MERCY) was conducted on patients who underwent laparoscopic or robotic right colectomy between 2014 and 2020. Classification tree approach was used to describe the extracorporeal (EA) or intracorporeal (IA) anastomosis choice. Mixed-model regressions were used to identify patient- and surgery-related factors predictive of postoperative outcomes. A questionnaire was used to evaluate the surgeons' perspectives.
Results:
The MERCY database comprised 1870 patients; 87.2% underwent laparoscopy, and 68.1% received an EA. A clear surgical trend was noted, with an increasing rate of IA and robotic procedures after 2017. EA represented 41% of anastomoses in centers equipped with a robotic surgical system. Mixed-model regressions (on 1088 patients) showed that age, sex, BMI, comorbidity, robotics, IA, and conversion to open surgery were predictors of surgical outcomes. In particular, IA was a predictor of a shorter time to regular diet and fewer surgical site infections. Based on the questionnaire, IA was the preferred over EA by 72% of surgeons.
Conclusion:
MIS continues to evolve, with an increasing number of IA being performed in the recent years and when using a robotic surgical system. Understanding the role of predictors of surgical outcomes may help surgeons personalize decision-making among the different MIS options to manage right colon cancer.
... Finally, we computed the Deyo version of the Carlson Comorbidity Index. [27][28][29][30] We defined the comorbidities using Read Codes in the CPRD recorded any time before study cohort entry, and prescriptions were defined using British National Formulary codes in the year before study cohort entry. ...
5‐α reductase inhibitors (5αRIs) are effective for the treatment of benign prostatic hyperplasia (BPH). However, 5αRIs could lower levels of haemoglobin, increasing the risk of anaemia. The objective of this study was to compare the rate of anaemia between new users of 5αRIs and α‐blockers in the UK.
Methods
We conducted a matched, active comparator, new‐user cohort study using the Clinical Practice Research Datalink. The study population consisted of men aged ≥40 years with incident BPH who initiated 5αRIs between 1998 and 2019 and were matched 1:1 on propensity score to new users of α‐blockers. Anaemia was defined by a measured haemoglobin <130 g/L. We used Cox proportional hazards models to estimate hazard ratios (HRs) and 95% confidence intervals (CIs) for anaemia.
Results
Our study cohort included 9429 new users of 5αRIs and 9429 matched new users of α‐blockers. Their median durations of follow‐up were 136 days (interquartile range: 54–336 d) and 77 days (interquartile range: 58–236 d), respectively. A total of 2865 5αRIs users and 2407 α‐blocker users developed incident anaemia, representing rates of 37.3 (95% CI: 33.6–41.3) and 42.0 (95% CI: 38.1–46.2) per 100 person‐years, respectively. The use of 5αRIs was not associated with an increased risk of anaemia compared to the use of α‐blockers (HR: 0.95, 95% CI: 0.90–1.00). Similarly, we did not observe an increased risk of mild, moderate, or severe anaemia.
Conclusion
The use of 5αRIs was not associated with an increased risk of anaemia compared to the use of α‐blockers among men with BPH.
... A post hoc analysis was conducted to determine the number of patients who received a first prescription for lisdexamfetamine before and after the date of approval by the FDA for the treatment of BED, stratified by presence or absence of comorbid attentiondeficit/hyperactivity disorder (ADHD) at baseline. Two indices of overall health were calculated: the Framingham Risk Score (FRS) [30] to assess the 10-year risk of cardiovascular disease and the Charlson Comorbidity Index (CCI) [31,32] to assess the 1-year risk of mortality (see Supplementary Table 1). PPVs with 95% CIs are reported for the final BED algorithms and for algorithm-identified behavioral therapies. ...
Objectives
Treatment for adults diagnosed with binge-eating disorder (BED) includes psychotherapy and/or pharmacotherapy and aims to reduce the frequency of binge-eating episodes and disordered eating, improve metabolic-related issues and reduce weight, and address mood symptoms. Data describing real-world treatment patterns are lacking; therefore, this study aims to characterize real-world treatment patterns among patients with BED.
Methods
This retrospective study identified adult patients with BED using natural language processing of clinical notes from the Optum electronic health record database from 2009 to 2015. Treatment patterns were examined during the 12 months preceding the BED recognition date and during a follow-up period after BED recognition (1–3 years for most patients).
Results
Among 1042 patients, 384 were categorized as the BED cohort and 658, who met less stringent criteria, were categorized as probable BED. In the BED cohort, mean ± SD age was 45.2 ± 13.4 years and 81.8% were women (probable BED, 45.9 ± 12.8 years, 80.2%). A greater percentage of patients in the BED cohort were prescribed pharmacotherapy (70.6% [probable BED, 66.9%]) than received/discussed psychotherapy (53.1% [probable BED, 39.2%]) at baseline. In the BED cohort, 54.4% of patients were prescribed antidepressants (probable BED, 52.4%), 25.3% stimulants (probable BED, 20.1%), and 34.4% nonspecific psychotherapy (probable BED, 24.6%) at baseline, with no substantive differences observed during follow-up. Low percentages of patients in the BED cohort received/discussed cognitive behavioral therapy at baseline (12.5% [probable BED, 9.0%) or during follow-up (13.0% [probable BED, 8.8%). Among patients with ≥1 psychotherapy visit, the mean ± SD number of visits in the BED cohort was 1.2 ± 5.9 at baseline (probable BED, 1.7 ± 7.3) and 2.2 ± 7.7 during follow-up (probable BED, 2.6 ± 7.7).
Conclusion
This cohort of patients with BED was treated more frequently with pharmacotherapy than psychotherapy. These data may help inform strategies for reducing differences between real-world treatment patterns and evidence-based recommendations.
... Dubious or incomplete data about these measures were complemented with the information of the participant's PCP, relatives or proxies. With the available participant's information, two experienced investigators calculated the scores in two objective health scales: the Carey scale and the Charlson-Romano comorbidity index [32,33]. The Carey scale is an UK primary care elderly patients (n = 653,780) scale validated to determined shortterm mortality. ...
Background
The causes of the dementia decrease in affluent countries are not well known but health amelioration could probably play a major role. Nevertheless, although many vascular and systemic disorders in adult life are well-known risk factors (RF) for dementia and Alzheimer disease (AD), health status is rarely considered as a single RF.
Aim
To analyse whether the health status and the self-perceived health (SPH) could be RF for dementia and AD and to discuss its biological basis.
Methods
We analysed different objective health measures and SPH as RF for dementia and AD incidence in 4569 participants of the NEDICES cohort by means of Cox-regression models. The mean follow-up period was 3.2 (range: 0.03–6.6) years.
Results
Ageing, low education, history of stroke, and “poor” SPH were the main RF for dementia and AD incidence, whereas physical activity was protective. “Poor” SPH had a hazard ratio = 1.66 (95% CI 1.17–2.46; p = 0.012) after controlling for different confounders.
Discussion
According to data from NEDICES cohort, SPH is a better predictor of dementia and AD than other more objective health status proxies. SPH should be considered a holistic and biologically rooted indicator of health status, which can predict future development of dementia and AD in older adults.
Conclusions
Our data indicate that it is worthwhile to include the SPH status as a RF in the studies of dementia and AD incidence and to explore the effect of its improvement in the evolution of this incidence.
... The adapted Charlson comorbidity score (0, 1, or ≥2) used Medicare claims from the year before cancer. 49,50 We included the cancer site and American Joint Committee on Cancer staging from SEER, and we ascertained the receipt of treatment (surgery, radiation therapy, and chemotherapy) in the year after diagnosis from Medicare claims. We included the presence or absence of an LTOT episode in the first year after diagnosis, the year in which the index LTOT episode started, and the months from 1 year after diagnosis to the start of the index LTOT episode by using Medicare Part D claims. ...
Background
Cancer survivors receive more long‐term opioid therapy (LTOT) than people without cancer, but the safety of LTOT prescribing is unknown.
Methods
Opioid‐naive adults aged ≥66 years who had been diagnosed in 2008‐2015 with breast, lung, head and neck, or colorectal cancer were identified with data from Surveillance, Epidemiology, and End Results cancer registries linked with Medicare claims. Survivors with 1 or more LTOT episodes (≥90 consecutive days) occurring ≥1 year after their cancer diagnosis and before censoring at hospice entry, another cancer diagnosis, 6 months before death, or December 2016 were included. The safety of prescribing during the first 90 days of the first LTOT episode was measured during follow‐up. As a positive safety indicator, the proportion of survivors with concurrent nonopioid pain management was measured. Indicators of less safe prescribing were the proportion of survivors with a high average daily opioid dose (≥90 morphine milligram equivalents) and the proportion of survivors with concurrent benzodiazepine dispensing. Multivariable logistic regression analyses were conducted to identify clinical predictors of each safety outcome.
Results
In all, 3628 cancer survivors received LTOT during follow‐up (median duration, 4.9 months; interquartile range, 3.5‐8.0 months). Seventy‐two percent of the survivors received multimodal pain management concurrently with LTOT. Eight percent of the survivors had high‐dose opioid prescriptions; 25% of the survivors received benzodiazepines during LTOT. Multivariable analyses identified variations in safety measures by multiple clinical factors, although none were consistently significant across outcomes.
Conclusions
To improve safe LTOT prescribing for survivors, efforts should focus on increasing multimodal pain management and reducing inappropriate benzodiazepine prescribing. Different clinical predictors of each outcome suggest different drivers of safe prescribing.
... Urbanisation was categorised into the following five levels, which are used specifically in Taiwan: Level 1 (highly urbanised area), Level 2 (moderately urbanised area), Level 3 (township area), Level 4 (subrural area) and Level 5 (rural area). The CCI was used to categorise the severity of medical comorbidities (assigned with scores of 0, 1 or ⩾2) according to the ICD-9-CM codes for primary and other diagnoses in the NHIRD (Charlson et al., 1987;Romano et al., 1993). The CCI comprised 17 medical illnesses within 1 year prior to the index date, namely, myocardial infarction, congestive heart failure, peripheral vascular disease, cerebrovascular disease, dementia, chronic pulmonary disease, connective tissue disease, ulcer disease, mild liver disease, diabetes, diabetes with end organ damage, hemiplegia, moderate or severe renal disease, any tumour, moderate or severe liver disease, metastatic solid tumour and acquired immune deficiency syndrome. ...
Objective
Medical comorbidities are prevalent in patients with bipolar disorder. Evaluating longitudinal trends of the incidence of medical illnesses enables implementation of early prevention strategies to reduce the high mortality rate in this at-risk population. However, the incidence risks of medical illnesses in the early stages of bipolar disorder remain unclear. This study investigated the incidence and 5-year trend of medical illnesses following bipolar disorder diagnosis.
Methods
We identified 11,884 patients aged 13–40 years who were newly diagnosed as having bipolar disorder during 1996–2012 and 47,536 age- and sex-matched controls (1:4 ratio) who represented the general population from Taiwan’s National Health Insurance Research Database. We estimated the prevalence and incidence of individual medical illnesses yearly across the first 5 years after the index date. The adjusted incidence rate ratio was calculated to compare the occurrence of specific medical illnesses each year between the bipolar disorder group and control group using the Poisson regression model.
Results
Apart from the prevalence, the adjusted incidence rate ratios of most medical illnesses were >1.00 across the first 5-year period after bipolar disorder diagnosis. Cerebrovascular diseases, ischaemic heart disease, congestive heart failure, other forms of heart disease, renal disease and human immunodeficiency virus infection exhibited the highest adjusted incidence rate ratios during the first year. Except for that of renal disease, the 5-year trends of the adjusted incidence rate ratios decreased for cerebrovascular diseases, cardiovascular diseases (e.g. ischaemic heart disease, other forms of heart disease, and vein and lymphatic disease), gastrointestinal diseases (e.g. chronic hepatic disease and ulcer disease) and communicable diseases (e.g. human immunodeficiency virus infection, upper respiratory tract infection and pneumonia).
Conclusion
Incidence risks of medical illnesses are increased in the first year after bipolar disorder diagnosis. Clinicians must carefully evaluate medical illnesses during this period because the mortality rates from medical illnesses are particularly high in people with bipolar disorder.
... Medical history variables included psychiatric illness, mechanical neck or back problems (excluding low back pain), mechanical low back pain, osteoarthritis, rheumatoid arthritis, diabetic neuropathy, peripheral neuropathy (excluding diabetic neuropathy), lumbar radiculopathy, alcohol dependence or harmful use, opioid use disorder and Romano comorbidity score (an index of the patient's comorbidities based on previous diagnoses). 16,17 (See Appendix 1, Tables S3 and S4, for diagnostic codes and definitions for chronic pain conditions.) [18][19][20][21][22][23][24][25][26] Prescription history variables included opioid substitution therapy (≥ 1 prescription), sedative and hypnotic medication (including benzodiazepines) (≥ 1 prescription), maximum daily opioid analgesic dose prescribed (≤ 50 mg of morphine equivalents [MME], > 50 to 90 MME, > 90 to 200 MME or > 200 MME) 27 and intensity of opioid analgesic use (60 to < 90 days' supply or ≥ 90 days' supply prescribed). ...
... Charlson Comorbidity Index (CCI) scores were calculated according to the Dartmouth-Manitoba adaptation. 24 Diagnosis codes were used to identify claims for fall/fracture, which was an outcome of interest given the relationship between vision loss and fall/ fracture. Prior anti-VEGF use was identified using a combination of Healthcare Common Procedure Coding System (HCPCS) codes and National Drug Codes (NDC) for bevacizumab, ranibizumab, and aflibercept; proxy rules were used to identify intravitreal bevacizumab use for nAMD, as bevacizumab has multiple indications (Supplementary Table 2, available in online article). ...
BACKGROUND: New treatment alternatives have revolutionized the management of nAMD. However, there is limited evidence on the clinical and economic burden of nAMD in commercially insured US patients. OBJECTIVES: To examine the clinical and economic burden in patients with nAMD by disease status in the commercially insured US patient population and to identify drivers of nAMD-related costs. METHODS: Patients with at least 1 International Classification of Diseases, 10th Revision Clinical Modification (ICD-10-CM) diagnosis for nAMD were identified from the IQVIA PharMetrics Plus database between April 2016 and August 2017 (index period). Patients had continuous enrollment for at least 6 months before and at least 12 months after the index date. Eye-level disease status was reported, along with intravitreal anti-VEGF treatment patterns. Health care resource utilization (HRU) (all-cause and nAMD-related) and direct health care costs were estimated over the 12 month follow-up period. Outcomes associated with falls and fractures were also assessed. Multivariate analysis identified drivers of annual nAMD-related outpatient costs among patients with anti-VEGF therapy. Incident patients (defined as those without an nAMD diagnosis 6 months prior to the index date) with at least 18 months of continuous enrollment after the index date were identified for a subset analysis to evaluate documented changes in disease status. RESULTS: A total of 6,076 patients with nAMD were identified for the prevalent cohort; 60.1%, 17.2%, and 5.9% had active CNV, inactive CNV, and inactive scar disease stage at index, respectively. The nAMD-related outpatient visit costs were roughly 4 and roughly 7 times higher, respectively, for the active CNV group (11,612]) compared with the inactive CNV (5,510]) and inactive scar (3,035]) groups (P < 0.0001). About 10% of prevalent patients had a fall/fracture claim over 12 months of follow-up. A total of 3,623 prevalent patients (59.6%) were eligible for the anti-VEGF treatment patterns analysis (mean [SD] duration of therapy = 7.7 [4.5] months; mean [SD] number of injections = 6.0 [3.7]). Qualified incident cases comprised 17.8% (n = 1,081) of the prevalent cohort. Approximately 20% of incident eyes with active CNV at baseline transitioned to inactive CNV. A total of 427 incident patients (39.5%) qualified for anti-VEGF treatment patterns analysis (mean [SD] duration of therapy = 6.2 [4.7] months, mean [SD] number of injections = 5.2 [3.5]). Significant drivers of total nAMD-related costs were the initial anti-VEGF agent and anti-VEGF injection frequency (P < 0.0001) in both prevalent and incident cohorts. CONCLUSIONS: The clinical and economic burden of nAMD treatment is substantial to the US healthcare system, where economic burden is higher among those with active CNV. Appropriate treatment may increase the duration of inactive disease periods and preserve visual acuity while lowering costs. DISCLOSURES: This study was funded by Allergan, an AbbVie Company. Allergan employees were involved in the study design, interpretation of data, writing of the manuscript, and the decision to submit for publication. Keyloun and Campbell are employees of Allergan. Multani, McGuiness, and Chen are employees of IQVIA, which received funding from Allergan for conducting the analysis. Almony and Shah-Manek have nothing to disclose.
... CCI was developed and validated in 1987 [3] to predict risk of death within 1 year of hospitalization and since then different versions that uses administrative data have been created [4]. CCI was first adapted to ICD-codes by Deyo in 1992 and Romani in 1993 [5,6]. Today, the most commonly used is the updated version by Quan et al. [4] that adapted CCI to be used with ICD-10 codes. ...
Background
Several comorbidity indices have been created to estimate and adjust for the burden of comorbidity. The objective of this systematic review was to evaluate and compare the ability of different comorbidity indices to predict mortality in an orthopedic setting.
Methods
A systematic search was conducted in Embase, MEDLINE, and Cochrane Library. The search were constructed around two primary focal points: a comorbidity index and orthopedics. The last search were performed on 13 June 2019. Eligibility criteria were participants with orthopedic conditions or who underwent an orthopedic procedure, a comparison between comorbidity indices that used administrative data, and reported mortality as outcome. Two independent reviewers screened the studies using Covidence. The area under the curve (AUC) was chosen as the primary effect estimate.
Results
Of the 5338 studies identified, 16 met the eligibility criteria. The predictive ability of the different comorbidity indices ranged from poor (AUC < 0.70) to excellent (AUC ≥ 0.90). The majority of the included studies compared the Elixhauser Comorbidity Index (ECI) and the Charlson Comorbidity Index (CCI). In-hospital mortality was reported in eight studies reporting AUC values ranging from 0.70 to 0.92 for ECI and 0.68 to 0.89 for CCI. AUC values were generally lower for all other time points ranging from 0.67 to 0.78. For 1-year mortality the overall effect size ranging from 0.67 to 0.77 for ECI and 0.69 to 0.77 for CCI.
Conclusion
The results of this review indicate that the ECI and CCI can equally be used to adjust for comorbidities when analyzing mortality in an orthopedic setting.
Trial registration
The protocol for this systematic review was registered on PROSPERO, the International Prospective Register of Systematic Reviews on 13 June 2019 and can be accessed through record ID 133,871.
... The Charlson index was developed to predict 1-year mortality risk by assigning empirically derived scores to 19 clinically relevant comorbidities [11]. The Romano index, the most commonly used modification of Charlson index, groups the comorbidities on the basis of the International Classification of Diseases (ICD) diagnosis coding system (9 th & 10 th 1 3 version) [12,13]. A scoring system introduced by Elixhauser et al. predicts in-hospital mortality and length of stay [14] and, because it takes into account more comorbidities, may be superior to the Charlson/Romano index in predicting risk in certain surgical cohorts [15,16]. ...
The American Society of Anesthesiologists (ASA) Physical Status Classification System has been used to assess pre-anesthesia comorbid conditions for over 60 years. However, the ASA Physical Status Classification System has been criticized for its subjective nature. In this study, we aimed to assess the correlation between the ASA physical status assignment and more objective measures of overall illness.
This is a single medical center, retrospective cohort study of adult patients who underwent surgery between November 2, 2017 and April 22, 2020. A multivariable ordinal logistic regression model was developed to examine the relationship between the ASA physical status and Elixhauser comorbidity groups. A secondary analysis was then conducted to evaluate the capability of the model to predict 30-day postoperative mortality.
A total of 56,820 cases meeting inclusion criteria were analyzed. Twenty-seven Elixhauser comorbidities were independently associated with ASA physical status. Older patient (adjusted odds ratio, 1.39 [per 10 years of age]; 95% CI 1.37 to 1.41), male patient (adjusted odds ratio, 1.24; 95% CI 1.20 to 1.29), higher body weight (adjusted odds ratio, 1.08 [per 10 kg]; 95% CI 1.07 to 1.09), and ASA emergency status (adjusted odds ratio, 2.11; 95% CI 2.00 to 2.23) were also independently associated with higher ASA physical status assignments. Furthermore, the model derived from the primary analysis was a better predictor of 30-day mortality than the models including either single ASA physical status or comorbidity indices in isolation (p < 0.001).
We found significant correlation between ASA physical status and 27 of the 31 Elixhauser comorbidities, as well other demographic characteristics. This demonstrates the reliability of ASA scoring and its potential ability to predict postoperative outcomes. Additionally, compared to ASA physical status and individual comorbidity indices, the derived model offered better predictive power in terms of short-term postoperative mortality.
Background
Epidemiological evidence on biosimilars’ real-world performance is limited. On July 18th, 2017, a biosimilar health policy was implemented in British Columbia (BC), Canada, mandating all patients initiating a new biologic medication to be prescribed a biosimilar (if/when available). Exploiting a policy change as a natural experiment, we assessed the real-world impact of biosimilar use for inflammatory arthritis (IA) on health resource utilization as a surrogate marker of real-world effectiveness and safety.
Methods
Using administrative health data, we identified all incident etanercept users for IA in BC with initiation dates between 2014 and 2020 (n = 3004) [63·6% female; mean (S.D.) age at IA disease diagnosis 52·5 (16·6) years]. Healthcare utilization over three years after initiation was assessed using outcomes including — physician visits (PV), all-cause hospitalizations (ACH), infection-related hospitalizations (IRH), length of hospital stays (LOS), and emergency room visits (ERV). Using regression discontinuity design, we compared healthcare utilization risk in patients initiating etanercept immediately before/after policy-change date, representing the intention-to-treat effect. Additionally, we estimated the complier average causal effect of biosimilar use with instrumental variable (IV) control function method.
Findings
Intention-to-treat analyses showed no significant impact of biosimilar policy implementation on PV, HOSP, IRH, LOS, or ERV, with respective adjusted RRs of 0·96 (95% CI: 0·82–1·12), 0·84 (95% CI: 0·49–1·44), 0·91 (95% CI: 0·21–3·86), 0·94 (95% CI: 0·41–2·15), and 0·91 (95% CI: 0·44–1·88). IV analyses indicated biosimilar use in routine settings did not significantly change healthcare utilization, compared to originator etanercept.
Interpretation
No significant impact of biosimilar policy or actual biosimilar use on healthcare utilization was observed, suggesting equivalent real-world effectiveness and safety of biosimilars to originators and no unintended consequences of the policy change.
Funding
CHIR and 10.13039/501100000038NSERC.
PURPOSE
Multiple myeloma (MM) is the most common plasma cell dyscrasia in the United States with notably significant health disparities. MM is preceded by an asymptomatic precursor monoclonal gammopathy of undetermined significance (MGUS). Studies have identified several risk factors for the progression of MGUS to MM; however, the relative contributions of these remain unknown. Particularly, understanding the contribution among those modifiable factors may inform MM prevention.
METHODS
This study quantified these contributions by estimating the adjusted population attributable fractions (aPAF) of modifiable risk factors for MM among the Veteran population with MGUS.
RESULTS
Among all evaluated risk factors, excess body mass index (BMI ≥25 kg/m ² ) was the leading factor (Black: aPAF=27.0%, 95% CI 19.3-33.9%; White: 27.1%, 95% CI 20.3-33.4%; All: aPAF=27.1%, 95% CI: 22.0-31.8%).
CONCLUSION
Our study highlights the potential for weight management as a key strategy in reducing the risk of progression to MM in Black and White patients diagnosed with MGUS.
Objective
Individuals with colorectal cancer (CRC) have an increased risk of mental disorders, yet mental healthcare utilisation has not been adequately examined. We evaluated mental healthcare utilisation and receipt of minimally adequate treatments for anxiety and/or depression among individuals with and without CRC.
Methods and analysis
We used administrative health databases from British Columbia, Canada, comprised of individuals with CRC and individuals without CRC, matched (1:1 ratio) on age, sex and incident mental disorder(s) (ie, occurring after CRC diagnosis/matched date). Primary outcomes were minimally adequate antidepressant pharmacotherapy (≥84 days’ supply) and psychological (≥4 services) treatment.
Results
Among individuals with CRC, 1462 had incident anxiety (mean age 64.6±12.5 years, 59.2% females), 4640 had incident depression (mean age 66.3±12.3 years, 51.2% females). Approximately one in four individuals with CRC were diagnosed with anxiety (23.4%) and/or depression (23.2%) in the first year after CRC diagnosis. Minimally adequate antidepressant pharmacotherapy (36.2%) and psychological treatment (15.9%) for anxiety were significantly lower in CRC patients than in those without CRC (pharmacotherapy adjusted OR (aOR) 0.74; 95% CI 0.61, 0.88; psychological treatment aOR 0.74; 95% CI 0.58, 0.95). Similar findings were observed for depression (pharmacotherapy aOR 0.81; 95% CI 0.74, 0.90). Among individuals with CRC, mental healthcare utilisation persisted up to 10 years post-mental disorder diagnosis.
Conclusions
Individuals with CRC receive less mental health treatment for anxiety and/or depression, compared with those without CRC. Findings raise awareness for the need for ongoing mental healthcare throughout and beyond CRC.
Importance
Semaglutide, a glucagonlike peptide-1 receptor agonist (GLP-1RA), has recently been implicated in cases of nonarteritic anterior ischemic optic neuropathy (NAION), raising safety concerns in the treatment of type 2 diabetes (T2D).
Objective
To investigate the potential association between semaglutide and NAION in the Observational Health Data Sciences and Informatics (OHDSI) network.
Design, Setting, and Participants
This was a retrospective study across 14 databases (6 administrative claims and 8 electronic health records). Included were adults with T2D taking semaglutide, other GLP-1RA (dulaglutide, exenatide), or non–GLP-1RA medications (empagliflozin, sitagliptin, glipizide) from December 1, 2017, to December 31, 2023. The incidence proportion and rate of NAION were calculated. Association between semaglutide and NAION was assessed using 2 approaches: an active-comparator cohort design comparing new users of semaglutide with those taking other GLP-1RAs and non–GLP-1RA drugs, and a self-controlled case-series (SCCS) analysis to compare individuals’ risks during exposure and nonexposure periods for each drug. The cohort design used propensity score–adjusted Cox proportional hazards models to estimate hazard ratios (HRs). The SCCS used conditional Poisson regression models to estimate incidence rate ratios (IRRs). Network-wide HR and IRR estimates were generated using a random-effects meta-analysis model.
Exposures
GLP-1RA and non–GLP-1RAs.
Main Outcomes and Measures
NAION under 2 alternative definitions based on diagnosis codes: one more inclusive and sensitive, the other more restrictive and specific.
Results
The study included 37.1 million individuals with T2D, including 810 390 new semaglutide users. Of the 43 620 new users of semaglutide in the Optum’s deidentified Clinformatics Data Mart Database, 24 473 (56%) were aged 50 to 69 years, and 26 699 (61%) were female. The incidence rate of NAION was 14.5 per 100 000 person-years among semaglutide users. The HR for NAION among new users of semaglutide was not different compared with that of the non–GLP-1RAs using the sensitive NAION definition—empagliflozin (HR, 1.44; 95% CI, 0.78-2.68; P = .12), sitagliptin (HR, 1.30; 95% CI, 0.56-3.01; P = .27), and glipizide (HR, 1.23; 95% CI, 0.66-2.28; P = .25). The risk was higher only compared with patients taking empagliflozin (HR, 2.27; 95% CI, 1.16-4.46; P = .02) using the specific definition. SCCS analysis of semaglutide exposure showed an increased risk of NAION (meta-analysis IRR, 1.32; 95% CI, 1.14-1.54; P < .001).
Conclusions and Relevance
Results of this study suggest a modest increase in the risk of NAION among individuals with T2D associated with semaglutide use, smaller than that previously reported, and warranting further investigation into the clinical implications of this association.
Objective
To quantify the comparative risk of thrombosis with thrombocytopenia syndrome or thromboembolic events associated with use of adenovirus based covid-19 vaccines versus mRNA based covid-19 vaccines.
Design
International network cohort study.
Setting
Routinely collected health data from contributing datasets in France, Germany, the Netherlands, Spain, the UK, and the US.
Participants
Adults (age ≥18 years) registered at any contributing database and who received at least one dose of a covid-19 vaccine (ChAdOx1-S (Oxford-AstraZeneca), BNT162b2 (Pfizer-BioNTech), mRNA-1273 (Moderna), or Ad26.COV2.S (Janssen/Johnson & Johnson)), from December 2020 to mid-2021.
Main outcome measures
Thrombosis with thrombocytopenia syndrome or venous or arterial thromboembolic events within the 28 days after covid-19 vaccination. Incidence rate ratios were estimated after propensity scores matching and were calibrated using negative control outcomes. Estimates specific to the database were pooled by use of random effects meta-analyses.
Results
Overall, 1 332 719 of 3 829 822 first dose ChAdOx1-S recipients were matched to 2 124 339 of 2 149 679 BNT162b2 recipients from Germany and the UK. Additionally, 762 517 of 772 678 people receiving Ad26.COV2.S were matched to 2 851 976 of 7 606 693 receiving BNT162b2 in Germany, Spain, and the US. All 628 164 Ad26.COV2.S recipients from the US were matched to 2 230 157 of 3 923 371 mRNA-1273 recipients. A total of 862 thrombocytopenia events were observed in the matched first dose ChAdOx1-S recipients from Germany and the UK, and 520 events after a first dose of BNT162b2. Comparing ChAdOx1-S with a first dose of BNT162b2 revealed an increased risk of thrombocytopenia (pooled calibrated incidence rate ratio 1.33 (95% confidence interval 1.18 to 1.50) and calibrated incidence rate difference of 1.18 (0.57 to 1.8) per 1000 person years). Additionally, a pooled calibrated incidence rate ratio of 2.26 (0.93 to 5.52) for venous thrombosis with thrombocytopenia syndrome was seen with Ad26.COV2.S compared with BNT162b2.
Conclusions
In this multinational study, a pooled 30% increased risk of thrombocytopenia after a first dose of the ChAdOx1-S vaccine was observed, as was a trend towards an increased risk of venous thrombosis with thrombocytopenia syndrome after Ad26.COV2.S compared with BNT162b2. Although rare, the observed risks after adenovirus based vaccines should be considered when planning further immunisation campaigns and future vaccine development.
Many mathematical models have been proposed to predict death following COVID-19; all started with comorbidity subsets for this still-little understood disease. Thus, we derived a novel predicted probability of death model (PDeathDx) upon all diagnostic codes documented in the Department of Veterans Affairs. We present the conceptual underpinnings and analytic approach in estimating the independent contribution of pre-existing conditions. This is the largest study to-date following patients with COVID-19 to predict mortality. Cases were identified with at least one positive nucleic acid amplification test (NAAT). Starting in 1997, we use diagnoses from the first time a patient sought care until 14 days before a positive NAAT. We demonstrate the clear advantage of using an unrestricted set of pre-existing conditions to model COVID-19 mortality, as models using conventional comorbidity indices often assign little weight or usually do not include some of the highest-risk conditions; the same is true of conditions associated with COVID-19 severity. Our findings suggest that it is risky to pick comorbidities for analysis without a systematic review of all those experienced by the cohort. Unlike conventional approaches, our comprehensive methodology provides the flexibility that has been advocated for comorbidity indices since 1993; such an approach can be readily adapted for other diseases and outcomes. With our comorbidity risk adjustment approach outperforming conventional indices for predicting COVID-19 mortality, it shows promise for predicting outcomes for other conditions of interest.
Rationale
The long-term natural history of asthma in terms of successive severe exacerbations and the influence of each exacerbation on the course of the disease is not well studied.
Objectives
To investigate the long-term natural history of asthma among patients who are hospitalized for asthma for the first time in terms of the risk of future severe exacerbations and heterogeneity in this risk across patients.
Methods
Using the administrative health databases of British Columbia, Canada (January 1, 1997 to March 31, 2016), we created an incident cohort of patients with at least one asthma exacerbation that required inpatient care. We estimated the 5-year cumulative incidence of severe exacerbations after successive numbers of previous events. We used a joint frailty model to investigate the extent of between-individual variability in exacerbation risk and the associations of each exacerbation with the rate of subsequent events. Analyses were conducted separately for pediatric (<14 years old) and adult (⩾14 years old) patients.
Results
Analyses were based on 3,039 pediatric (mean age at baseline, 6.4; 35% female) and 5,442 (mean age at baseline, 50.8; 68% female) adult patients. The 5-year rates of severe exacerbations after the first three events were 0.16, 0.29, and 0.35 for the pediatric group, and 0.14, 0.33, and 0.49 for the adult group. Both groups exhibited substantial variability in patient-specific risks of exacerbation: the mid-95% interval of 5-year risk of experiencing a severe exacerbation ranged from 11% to 24% in pediatric patients and from 8% to 40% in adult patients. After controlling for potential confounders, the first follow-up exacerbation was associated with an increase of 79% (95% confidence interval [CI], 11–189%) in the rate of subsequent events in the pediatric group, whereas this increase was 188% (95% CI, 35–515%) for the adult group. The effects of subsequent exacerbations were not statistically significant.
Conclusions
After the first severe exacerbation, the risk of subsequent events is substantially different among patients. The number of previous severe exacerbations carries nuanced prognostic information about future risk. Our results suggest that severe exacerbations in the early course of asthma detrimentally affect the course of the disease and risk of subsequent exacerbations.
"Time burden" (time required during treatment) is relevant when choosing a local therapy option for early-stage breast cancer but has not been rigorously studied. We compared the time burden for three common local therapies for breast cancer: (1) lumpectomy plus whole-breast irradiation (Lump+WBI), (2) mastectomy without radiation or reconstruction (Mast alone), and (3) mastectomy without radiation but with reconstruction (Mast+Recon).
Methods:
Using the MarketScan database, we identified 35,406 breast cancer patients treated from 2000 to 2011 with these local therapies. We quantified the total time burden as the sum of inpatient days (inpatient-days), outpatient days excluding radiation fractions (outpatient-days), and radiation fractions (radiation-days) in the first two years postdiagnosis. Multivariable regression evaluated the effect of local therapy on inpatient-days and outpatient-days adjusted for patient and treatment covariates.
Results:
Adjusted mean number of inpatient-days was 1.0 for Lump+WBI, 2.0 for Mast alone, and 3.1 for Mast+Recon (P < 0.001). Adjusted mean number of outpatient-days was 42.9 for Lump+WBI, 42.2 for Mast alone, and 45.8 for Mast+Recon (P < 0.001). The mean number of radiation-days for Lump+WBI was 32.4. Compared with Mast+Recon (48.9 days), total adjusted time burden was 4.7 days shorter for Mast alone (44.2 days) and 27.4 days longer for Lump+WBI (76.3 days). However, use of a 15 fraction WBI regimen would reduce the time burden differential between Lump+WBI and Mast+Recon to just 10.0 days.
Conclusions:
Although Mast+Recon confers the highest inpatient and outpatient time burden, Lump+WBI carries the highest total time burden. Increased use of hypofractionation will reduce the total time burden for Lump+WBI.
Background
Diffuse large B cell lymphoma (DLBCL) is an aggressive disease. The first‐line treatment is well defined in young patients; however, in oldest old patients treatment remains unclear.
Objectives
To investigate the impact of therapeutics management and geriatric evaluation on survival in aged patients with DLBCL.
Methods
We performed a systematic review of PubMed and COCHRANE databases of published report on elderly patients (median age 80 and above) with DLBCL, from January 2002 to January 2020.
Results
We included 32 studies (6 prospective and 26 retrospective). Patients treated with anthracyclines‐containing chemoimmunotherapy had a 2‐year overall survival (OS) of 59%–74.3% in prospective studies and 48.1–64.6% in retrospective studies. With less intensive treatment without anthracyclines, 2‐year OS was 28%–53%. Without specific treatment, median OS was 2 months. History of falls and severe comorbidities were associated with a decreased survival.
Conclusions
Chemoimmunotherapy with anthracyclines increases survival in selected very elderly patients in comparison with less intensive regimen. Geriatric assessment, in particular altered mobility disorders and severe comorbidities, is predictive of survival and should be associated with the therapeutic decision. More comparative studies are needed to guide the management of frailer patients.
Comorbidity indices currently used to estimate negative postoperative outcomes in orthopaedic surgery were originally developed among non-orthopaedic patient populations. While current indices were initially intended to predict short-term mortality, they have since been used for other purposes as well. As the rate of hip and knee arthroplasty steadily rises, understanding the magnitude of the effect of comorbid disease on postoperative outcomes has become increasingly more important. Currently, the ASA classification is the most commonly used comorbidity measure and is systematically recorded by the majority of national arthroplasty registries. Consideration should be given to developing an updated, standardized approach for comorbidity assessment and reporting in orthopaedic surgery, especially within the setting of elective hip and knee arthroplasty.
Cite this article: EFORT Open Rev 2021;6:629-640. DOI: 10.1302/2058-5241.6.200124
The objective of this study was to develop a prospectively applicable method for classifying comorbid conditions which might alter the risk of mortality for use in longitudinal studies. A weighted index that takes into account the number and the seriousness of comorbid disease was developed in a cohort of 559 medical patients. The 1-yr mortality rates for the different scores were: "0", 12% (181); "1-2", 26% (225); "3-4", 52% (71); and "greater than or equal to 5", 85% (82). The index was tested for its ability to predict risk of death from comorbid disease in the second cohort of 685 patients during a 10-yr follow-up. The percent of patients who died of comorbid disease for the different scores were: "0", 8% (588); "1", 25% (54); "2", 48% (25); "greater than or equal to 3", 59% (18). With each increased level of the comorbidity index, there were stepwise increases in the cumulative mortality attributable to comorbid disease (log rank chi 2 = 165; p less than 0.0001). In this longer follow-up, age was also a predictor of mortality (p less than 0.001). The new index performed similarly to a previous system devised by Kaplan and Feinstein. The method of classifying comorbidity provides a simple, readily applicable and valid method of estimating risk of death from comorbid disease for use in longitudinal studies. Further work in larger populations is still required to refine the approach because the number of patients with any given condition in this study was relatively small.
Analytic procedures suitable for the study of human disease are scattered throughout the statistical and epidemiologic literature. Explanations of their properties are frequently presented in mathematical and theoretical language. This text provides a clear understanding of the statistical methods that are widely used in epidemiologic research without depending on advanced mathematical or statistical theory. By applying these methods to actual data, this book reveals the strengths and weaknesses of each analytic approach. The book combines techniques from the fields of statistics, biostatistics, demography, and epidemiology to present an overview that does not require computational details of the statistical techniques described. Throughout, the text contains illuminating discussions with new elements for this edition, including the analysis of multi-level categorical data and simple, intuitive arguments that exponential survival times cause the hazard function to be constant. There are also new applied examples to illustrate such topics as the pitfalls of proportional mortality data, the analysis of matched pair categorical data, and the age-adjustment of mortality rates based on statistical models. The most important new feature is a chapter on Poisson regression analysis. This essential statistical tool permits the multivariable analysis of rates, probabilities, and counts.
Objective.
—A prospective regional study was conducted to determine if the observed differences in in-hospital mortality rates associated with coronary artery bypass grafting (CABG) are solely the result of differences in patient case mix.
Administrative databases are increasingly used for studying outcomes of medical care. Valid inferences from such data require the ability to account for disease severity and comorbid conditions. We adapted a clinical comorbidity index, designed for use with medical records, for research relying on International Classification of Diseases (ICD-9-CM) diagnosis and procedure codes. The association of this adapted index with health outcomes and resource use was then examined with a sample of Medicare beneficiaries who underwent lumbar spine surgery in 1985 (n = 27,111). The index was associated in the expected direction with postoperative complications, mortality, blood transfusion, discharge to nursing home, length of hospital stay, and hospital charges. These associations were observed whether the index incorporated data from multiple hospitalizations over a year's time, or just from the index surgical admission. They also persisted after controlling for patient age. We conclude that the adapted comorbidity index will be useful in studies of disease outcome and resource use employing administrative databases.
Percutaneous transluminal angioplasty has been adopted widely as a treatment for patients with peripheral vascular disease of the lower extremities. However, the effect of this procedure on the overall management of peripheral vascular disease and on the outcomes of patients has not been clearly delineated. In particular, it is not known whether angioplasty has replaced other treatments for peripheral vascular disease.
To assess the extent to which angioplasty is used and the associated changes in the surgical management of peripheral vascular disease of the lower extremities, we used data on hospital discharges in Maryland to identify all angioplasty procedures, peripheral bypass operations, and lower-extremity amputations performed for peripheral vascular disease in Maryland hospitals between 1979 and 1989.
We estimated that from 1979 to 1989 the annual rate of percutaneous transluminal angioplasty for peripheral vascular disease of the lower extremities, adjusted for age and sex, rose from 1 to 24 per 100,000 Maryland residents (P less than 0.0001 by linear regression). Despite this increase in the use of angioplasty, the adjusted annual rate of peripheral bypass surgery also rose substantially, from 32 to 65 per 100,000 (P less than 0.001), whereas the adjusted annual rate of lower-extremity amputation remained stable at about 30 per 100,000. Total charges for hospitalizations during which a peripheral revascularization procedure was performed increased from 30.5 million in 1989.
In Maryland, the adoption of percutaneous transluminal angioplasty for peripheral vascular disease of the lower extremities has been associated with an increase in the use of peripheral bypass surgery and with no decline in lower-extremity amputations. These results could be due to increased diagnosis of peripheral vascular disease, expanded indications for procedural interventions, or an increased number of repeat procedures performed in patients with peripheral vascular disease of the lower extremities.
How well can hospital discharge abstracts be used to estimate patient health status? This paper compares information on comorbidity obtained from hospital discharge abstracts for patients undergoing prostatectomy or cholecystectomy at a Winnipeg teaching hospital with clinical data on preoperative medical conditions prospectively collected during an Anesthesia Follow-up study. The diagnostic information on cardiovascular disease, respiratory disease, and metabolic disorders showed considerable agreement, ranging from 65 to over 90% correspondence across the two data sets. Certain conditions noted by the anesthesiologist were often absent from the claims data; cardiovascular disease was recorded in the clinical data but absent from the claims for 31% of prostatectomy and 17% of cholecystectomy cases. Such patients were less likely to have been assigned a high score on the ASA Physical Status measure or to have high-risk diagnoses on the hospital file. Similar findings resulted from comparing the two sources in their ability to predict such adverse outcomes as mortality and readmission to hospital: the anesthesia file generally included less serious comorbidity.
A prospective regional study was conducted to determine if the observed differences in in-hospital mortality rates associated with coronary artery bypass grafting (CABG) are solely the result of differences in patient case mix. DESIGN-Regional prospective cohort study. Data including patient demographic and historical data, body surface area, cardiac catheterization results, priority of surgery, comorbidity, and status at hospital discharge were collected. This study presents data for 3055 CABG patients between July 1, 1987, and April 15, 1989.
This study includes data from all surgeons performing cardiothoracic surgery in Maine, New Hampshire, and Vermont; the data were collected from five regional medical centers.
Data were collected from all consecutive isolated CABG surgery patients during the study period.
Crude and adjusted in-hospital mortality rates associated with CABG.
The overall crude in-hospital mortality rate for isolated CABG was 4.3%. The rate varied among centers (range, 3.1% to 6.3%) and among surgeons (range, 1.9% to 9.2%). Predictors of in-hospital mortality included increased age, female gender, small body surface area, greater comorbidity, reoperation, poorer cardiac function as indicated by a lower ejection fraction, increased left ventricular end diastolic pressure and emergent or urgent surgery. After adjusting for the effects of potentially confounding variables, substantial and statistically significant variability was observed among medical centers (P = .021) and among surgeons (P = .025).
We conclude that the observed differences in in-hospital mortality rates among institutions and among surgeons in northern New England are not solely the result of differences in case mix as described by these variables and may reflect differences in currently unknown aspects of patient care. Understanding this variation requires a detailed understanding of the processes of care.
In clinical practice, physicians informally address patient characteristics such as illness severity, comorbidity, functional status and stability when considering prognosis and diagnostic or therapeutic interventions. These same attributes, more formally measured, have been used as measures of casemix in clinical research to classify patients into similar risk strata. To determine whether physician estimates of illness severity, function status and stability were predictive of morbidity, mortality, length of stay and average daily ancillary charges, a cohort of 604 patients was studied. The predictive ability of the patient characteristics were found to be outcome specific. Illness severity was a significant predictor of in-hospital morbidity and mortality, length of stay and charges (p less than 0.001). Functional status was predictive of in-hospital morbidity and mortality, 1 year mortality, length of stay and charges (p less than 0.01). Physician estimates of stability were predictive only of morbidity (p less than 0.01) and comorbidity was only a preditor of 1 year mortality (p less than 0.001). Stratifying patients on the basis of specific clinical characteristics determined at the time of admission will be useful in studies of patient outcomes and resource utilization.
Per capita hospital expenditures in the United States exceed those in Canada, but little research has examined differences in outcomes. We used insurance databases to compare postsurgical mortality for 11 specific surgical procedures, both before and after adjustment for case mix, among residents of New England and Manitoba who were over 65 years of age. For low- and moderate-risk procedures, 30-day mortality rates were similar in both regions, but 6-month mortality rates were lower in Manitoba. For the two high-risk procedures, concurrent coronary bypass/valve replacement and hip fracture repair, both 30-day and 6-month mortality rates were lower in New England. Although no consistent pattern favoring New England for cardiovascular surgery was found, the increased mortality following hip fracture in Manitoba was found for all types of repair and all age groups. We conclude that for low- and moderate-risk procedures, the higher hospital expenditures in New England were not associated with lower perioperative mortality rates.
Claims-based indices of comorbidity and severity, as well as other measures derived from routinely collected administrative data, are developed and tested. The extent to which risk adjustments using claims can be improved by adding information from one well-known measure based on chart review and patient examination (the American Society of Anesthesiologists' (ASA) Physical Status score) is also examined. Readmissions and mortality after three common surgical procedures are the outcomes studied using multiple logistic regression. Claims-based measures of comorbidity, derived both from hospital discharge abstracts at the time of surgery and from hospitalizations in the 6 months before surgery, provided reasonably good predictions of postsurgical readmissions and mortality. In the most complete logistic regression models, the Somers' Dyx measure of fit (a rank correlation coefficient) ranged from 0.23 to 0.38 for readmissions and from 0.46 to 0.72 for mortality. In 5 out of 6 cases, these predictions were not improved by including the prospectively-collected ASA Physical Status score. Such difficulties in improving risk adjustment by more intensive data collection are discussed in terms of their research implications.
A system of scientific classification should have a suitable basic axis of organization, standardized names, clearly specified operational criteria, and multiaxial arrangements for citing important attributes beyond those included in the basic axis. During the past century, the main nosologic system for identifying human ailments has been the International Classification of Diseases (ICD), which has a well-organized and well-accepted nomenclature, but which lacks operational criteria and an appropriate multiaxial pattern. Two new systems of classification during the past two decades are intended for other purposes and would not be satisfactory as nosologic substitutes. The Problem-Oriented Record (POR) does not have a standardized nomenclature or criteria; and the Diagnosis-Related Group (DRG) approach was organized mainly for fiscal goals. As the basic taxonomy used for classifying human ailments, the ICD needs substantial improvement to fulfill its scientific role in statistics for the occurrence and treatment of disease.
Health services researchers rely heavily on administrative data bases, but incomplete or incorrect coding may bias risk models based on administrative data. The best method for validating administrative data is to collect detailed information about the same cases from independent sources, but this approach may be too costly or technically difficult. We used data on coronary artery bypass surgery from four sites (Duke University; Minneapolis--St Paul; California; and Manitoba) to demonstrate an alternative approach for assessing diagnostic coding and to explore the implications of miscoding. The first two sites have clinical data; the second two have administrative data. The prevalences of 14 comorbidities and the associated risk ratios for short-term mortality were compared across data sets. Some comorbidities could not be precisely mapped to ICD-9-CM. Chronic or asymptomatic conditions such as mitral insufficiency, cardiomegaly, previous myocardial infarction, tobacco use, and hyperlipidemia were far less prevalent in administrative data than in clinical data. The prevalence of diabetes, unstable angina, and congestive heart failure were similar in administrative and clinical data. Estimates of relative risk derived from clinical data equalled or surpassed those derived from administrative data for all conditions. Hospitals should be encouraged to improve reporting of coexisting conditions on discharge abstracts and claims. In the meantime, researchers using administrative data should assess the vulnerability of their risk models to bias caused by selective underreporting.
ICD-9-CM Coding Handbook With Amvers
- F Brown
Brown F. ICD-9-CM Coding Handbook With Amvers.
Chicago, IL: American Hospital Publishing; 1979.
The risk of adjustment: editorial
- H R Rubin
- A W Wu
Rubin HR, Wu AW. The risk of adjustment: editorial.
Med Care 1992; 30: 973-975.
MD: US Department of Health and Human Services
- Rockville
Rockville, MD: US Department of Health and Human
Services; 1992: AHCPR Publ. No. 92-0056.
Risk adjustment for medical outcome studies
- L I Iezzoni
Iezzoni LI. Risk adjustment for medical outcome
studies. In Grady ML, Schwartz H, Eds. Medical
Treatment Efkctiveness
Research Data Methods.