Modeling zero-modified count and semicontinuous data in health services research Part 1: background and overview

Respiratory drive heterogeneity associated with systemic inflammation and vascular permeability in acute respiratory distress syndrome

BACKGROUND

In acute respiratory distress syndrome (ARDS), respiratory drive often differs among patients with similar clinical characteristics. Readily observable factors like acid-base state, oxygenation, mechanics, and sedation depth do not fully explain drive heterogeneity. This study evaluated the relationship of systemic inflammation and vascular permeability markers with respiratory drive and clinical outcomes in ARDS.

METHODS

ARDS patients enrolled in the multicenter EPVent-2 trial with requisite data and plasma biomarkers were included. Neuromuscular blockade recipients were excluded. Respiratory drive was measured as PES0.1, the change in esophageal pressure during the first 0.1 s of inspiratory effort. Plasma angiopoietin-2, interleukin-6, and interleukin-8 were measured concomitantly, and 60-day clinical outcomes evaluated.

RESULTS

54.8% of 124 included patients had detectable respiratory drive (PES0.1 range of 0-5.1 cm H2O). Angiopoietin-2 and interleukin-8, but not interleukin-6, were associated with respiratory drive independently of acid-base, oxygenation, respiratory mechanics, and sedation depth. Sedation depth was not significantly associated with PES0.1 in an unadjusted model, or after adjusting for mechanics and chemoreceptor input. However, upon adding angiopoietin-2, interleukin-6, or interleukin-8 to models, lighter sedation was significantly associated with higher PES0.1. Risk of death was less with moderate drive (PES0.1 of 0.5-2.9 cm H2O) compared to either lower drive (hazard ratio 1.58, 95% CI 0.82-3.05) or higher drive (2.63, 95% CI 1.21-5.70) (p = 0.049).

CONCLUSIONS

Among patients with ARDS, systemic inflammatory and vascular permeability markers were independently associated with higher respiratory drive. The heterogeneous response of respiratory drive to varying sedation depth may be explained in part by differences in inflammation and vascular permeability.

An implicit measure of growth mindset uniquely predicts post-failure learning behavior

Research on implicit theories of intelligence (a.k.a. intelligence mindset) has shown that endorsing a stronger growth mindset (the belief that intelligence can be improved) is adaptive in the face of difficulties. Although the theory presumes implicit processes (i.e., unaware beliefs, guiding behaviors and actions automatically), the concept is typically assessed with self-reports. In this project we brought together research on intelligence mindset with research on implicit social cognition. Harnessing recent innovations from research on implicit measures, we assessed intelligence mindsets on an implicit level with a mousetracking Propositional Evaluation Paradigm. This measure captures the spontaneous truth evaluation of growth- and fixed-mindset statements to tap into implicit beliefs. In two preregistered laboratory studies (N = 184; N = 193), we found that implicitly measured growth mindsets predicted learning engagement after an experience of failure above and beyond the explicitly measured growth mindset. Our results suggest that implicit and explicit aspects of intelligence mindsets must be differentiated. People might be in a different mindset when making learning-related decisions under optimal conditions (i.e., with ample time and capacity) or under suboptimal conditions (i.e., when time pressure is high). This advancement in the understanding of implicit theories of intelligence is accompanied with substantial implications for theory and practice.

An Improved Statistical Modeling Approach to Individual Anticholinergic Drug Use Trend Analysis

Anticholinergic (AC) drugs are commonly prescribed to older adults for treating diseases and chronic conditions, such as chronic obstructive pulmonary disease, urinary incontinence, gastrointestinal disorder, or simply pain and allergy. The high prevalence of AC drug use can have a detrimental effect on the mental health of older adults. We aim to improve the prediction of future trends of AC drug use at the individual level, with pharmacy refill data. The individual drug use data presents challenges in the modeling, such as data being discrete-valued with excess zeros and having significant unobserved heterogeneity in the trend pattern. To address these challenges, we propose a statistical model of hierarchical structure and an EM scheme for the model parameter estimation. We evaluate the proposed modeling approach through a numerical study with synthetic data and a case study with real-world pharmacy refill data. The simulation study show that our analysis method outperforms the existing ones (e.g., reducing MSE significantly), particularly in terms of accurately predicting the trend pattern. The real-world case study further verifies the out-performance and demonstrate the advantageous features of our method. We expect the prediction tool developed based on our study can assist pharmacists' decision on initiating or strengthening behavioral interventions with the hope of discontinuing AC drug misuse.

Diagnostics for regression models with semicontinuous outcomes

Semicontinuous outcomes commonly arise in a wide variety of fields, such as insurance claims, healthcare expenditures, rainfall amounts, and alcohol consumption. Regression models, including Tobit, Tweedie, and two-part models, are widely employed to understand the relationship between semicontinuous outcomes and covariates. Given the potential detrimental consequences of model misspecification, after fitting a regression model, it is of prime importance to check the adequacy of the model. However, due to the point mass at zero, standard diagnostic tools for regression models (eg, deviance and Pearson residuals) are not informative for semicontinuous data. To bridge this gap, we propose a new type of residuals for semicontinuous outcomes that is applicable to general regression models. Under the correctly specified model, the proposed residuals converge to being uniformly distributed, and when the model is misspecified, they significantly depart from this pattern. In addition to in-sample validation, the proposed methodology can also be employed to evaluate predictive distributions. We demonstrate the effectiveness of the proposed tool using health expenditure data from the US Medical Expenditure Panel Survey.

Analysis of factors associated with the use of Korean medicine after spinal surgery using a nationwide database in Korea

Many patients in Korea use Korean Medicine (KM) after spine surgery, but related research is lacking. Therefore, this retrospective cohort study aimed to analyze factors affecting the use and costs of KM using nationally representative data from the National Health Insurance Service-National Sample Cohort, South Korea. Patients who underwent spinal surgery for spinal diseases from 2011 to 2014 were followed up for 5 years, and their medical care was described. The association between patient and spinal surgery characteristics and the use of KM was analyzed. A two-part model was used to analyze factors affecting the use of KM in patients undergoing spinal surgery. Of 11,802 patients who underwent spinal surgery, 11,367 who met the inclusion criteria were included. Overall, 55.5% were female, 32.3% were aged ≥ 70 years, and 50.2% received KM treatment during the follow-up period. Open discectomy was the most common surgical procedure performed (58.6%), and 40.2% of surgeries were performed because of lumbar disc disorder. Female sex, older age, high Charlson Comorbidity Index score, and use of KM before surgery were associated with increased KM use and expenditure after surgery. In conclusion, patient characteristics, rather than surgical characteristics, appeared to be more strongly associated with the use of KM after surgery, particularly prior experience with KM use. This study is significant in that it analyzed the entire spine surgery to provide a comprehensive view of the use of KM after spine surgery and analyzed the impact of various factors related patients and surgical characteristics on KM use. The results of this study may be useful to patients with spinal diseases, clinicians, and policymakers.

Association between multimorbidity and informal long-term care use in China: a nationwide cohort study

BACKGROUND

The impact of multimorbidity on long-term care (LTC) use is understudied, despite its well-documented negative effects on functional disabilities. The current study aims to assess the association between multimorbidity and informal LTC use in China. We also explored the socioeconomic and regional disparities.

METHODS

The study included 10,831 community-dwelling respondents aged 45 years and older from the China Health and Retirement Longitudinal Study in 2011, 2015, and 2018 for analysis. We used a two-part model with random effects to estimate the association between multimorbidity and informal LTC use. Heterogeneity of the association by socioeconomic position (education and income) and region was explored via a subgroup analysis. We further converted the change of informal LTC hours associated with multimorbidity into monetary value and calculated the 95% uncertainty interval (UI).

RESULTS

The reported prevalence of multimorbidity was 60·0% (95% CI: 58·9%, 61·2%) in 2018. We found multimorbidity was associated with an increased likelihood of receiving informal LTC (OR = 2·13; 95% CI: 1·97, 2·30) and more hours of informal LTC received (IRR = 1·20; 95% CI: 1·06, 1·37), ceteris paribus. Participants in the highest income quintile received more hours of informal LTC care (IRR = 1·62; 95% CI: 1·31, 1·99). The estimated monetary value of increased informal LTC hours among participants with multimorbidity was equivalent to 3·7% (95% UI: 2·2%, 5·4%) of China's GDP in 2018.

CONCLUSION

Our findings substantiate the threat of multimorbidity to LTC burden. It is imperative to strengthen LTC services provision, especially among older adults with multimorbidity and ensure equal access among those with lower income.

The cost burden of metastatic prostate cancer in the US populations covered by employer-sponsored health insurance

BACKGROUND

Recent advancements in the clinical management of metastatic prostate cancer include several costly therapies and diagnostic tests. The objective of this study was to provide updated information on the cost to payers attributable to metastatic prostate cancer among men aged 18 to 64 years with employer-sponsored health plans and men aged 18 years or older covered by employer-sponsored Medicare supplement insurance.

METHODS

By using Merative MarketScan commercial and Medicare supplemental data for 2009-2019, the authors calculated differences in spending between men with metastatic prostate cancer and their matched, prostate cancer-free controls, adjusting for age, enrollment length, comorbidities, and inflation to 2019 US dollars.

RESULTS

The authors compared 9011 patients who had metastatic prostate cancer and were covered by commercial insurance plans with a group of 44,934 matched controls and also compared 17,899 patients who had metastatic prostate cancer and were covered by employer-sponsored Medicare supplement plans with a group of 87,884 matched controls. The mean age of patients with metastatic prostate cancer was 58.5 years in the commercial samples and 77.8 years in the Medicare supplement samples. Annual spending attributable to metastatic prostate cancer was $55,949 per person-year (95% confidence interval [CI], $54,074-$57,825 per person-year) in the commercial population and $43,682 per person-year (95% CI, $42,022-$45,342 per person-year) in the population covered by Medicare supplement plans, both in 2019 US dollars.

CONCLUSIONS

The cost burden attributable to metastatic prostate cancer exceeds $55,000 per person-year among men with employer-sponsored health insurance and $43,000 among those covered by employer-sponsored Medicare supplement plans. These estimates can improve the precision of value assessments of clinical and policy approaches to the prevention, screening, and treatment of prostate cancer in the United States.

Aspergillus colonisation in severe community-acquired pneumonia: not just a mere colonisation

In patients with severe community-acquired pneumonia, detection of Aspergillus is associated with a mortality rate surpassing 50%, irrespective of whether it is defined as invasion or colonisation https://bit.ly/46PMk1f.

Geographical variation in the use of private health insurance in a predominantly publicly-funded system

We provide evidence of geographical variations in the use of private health insurance (PHI) in Italy. Our study offers an original contribution, using a 2016 dataset on the use of PHI amongst a population of more than 200,000 employees of a major company. The average claim per enrolee was €925, representing approximately 50% of public health expenditure per capita, primarily for dental care (27.2%), specialist outpatient services (26.3%) and inpatient care (25.2%). Residents in northern regions and metropolitan areas respectively claimed reimbursements for €164 and €483 more than those in southern regions and in non-metropolitan areas. Both supply and demand factors can explain these large geographical differences. The study suggests the urgency for policymakers to address the considerable disparities in the Italian healthcare system, revealing the overall social, cultural and economic conditions that shape the demand for healthcare.

Characterizing the Prevalence of Obesity Misinformation, Factual Content, Stigma, and Positivity on the Social Media Platform Reddit Between 2011 and 2019: Infodemiology Study

BACKGROUND

Reddit is a popular social media platform that has faced scrutiny for inflammatory language against those with obesity, yet there has been no comprehensive analysis of its obesity-related content.

OBJECTIVE

We aimed to quantify the presence of 4 types of obesity-related content on Reddit (misinformation, facts, stigma, and positivity) and identify psycholinguistic features that may be enriched within each one.

METHODS

All sentences (N=764,179) containing "obese" or "obesity" from top-level comments (n=689,447) made on non-age-restricted subreddits (ie, smaller communities within Reddit) between 2011 and 2019 that contained one of a series of keywords were evaluated. Four types of common natural language processing features were extracted: bigram term frequency-inverse document frequency, word embeddings derived from Bidirectional Encoder Representations from Transformers, sentiment from the Valence Aware Dictionary for Sentiment Reasoning, and psycholinguistic features from the Linguistic Inquiry and Word Count Program. These features were used to train an Extreme Gradient Boosting machine learning classifier to label each sentence as 1 of the 4 content categories or other. Two-part hurdle models for semicontinuous data (which use logistic regression to assess the odds of a 0 result and linear regression for continuous data) were used to evaluate whether select psycholinguistic features presented differently in misinformation (compared with facts) or stigma (compared with positivity).

RESULTS

After removing ambiguous sentences, 0.47% (3610/764,179) of the sentences were labeled as misinformation, 1.88% (14,366/764,179) were labeled as stigma, 1.94% (14,799/764,179) were labeled as positivity, and 8.93% (68,276/764,179) were labeled as facts. Each category had markers that distinguished it from other categories within the data as well as an external corpus. For example, misinformation had a higher average percent of negations (β=3.71, 95% CI 3.53-3.90; P<.001) but a lower average number of words >6 letters (β=-1.47, 95% CI -1.85 to -1.10; P<.001) relative to facts. Stigma had a higher proportion of swear words (β=1.83, 95% CI 1.62-2.04; P<.001) but a lower proportion of first-person singular pronouns (β=-5.30, 95% CI -5.44 to -5.16; P<.001) relative to positivity.

CONCLUSIONS

There are distinct psycholinguistic properties between types of obesity-related content on Reddit that can be leveraged to rapidly identify deleterious content with minimal human intervention and provide insights into how the Reddit population perceives patients with obesity. Future work should assess whether these properties are shared across languages and other social media platforms.

Utilization and out-of-pocket expenses of primary care among the multimorbid elderly in China: A two-part model with nationally representative data

Background

Multimorbidity has become an essential public health issue that threatens human health and leads to an increased disease burden. Primary care is the prevention and management of multimorbidity by providing continuous, comprehensive patient-centered services. Therefore, the study aimed to investigate the determinants of primary care utilization and out-of-pocket expenses (OOPE) among multimorbid elderly to promote rational utilization of primary care and reduce avoidable economic burdens.

Methods

The study used data from CHARLS 2015 and 2018, which included a total of 4,384 multimorbid elderly aged 60 and above. Guided by Grossman theory, determinants such as education, gender, marriage, household economy, and so on were included in this study. A two-part model was applied to evaluate primary care utilization and OOPE intensity in multimorbid populations. And the robustness testing was performed to verify research results.

Results

Primary care visits rate and OOPE indicated a decline from 2015 to 2018. Concerning primary outpatient care, the elderly who were female (OR = 1.51, P < 0.001), married (OR = 1.24, P < 0.05), living in rural areas (OR = 1.77, P < 0.001) and with poor self-rated health (OR = 2.23, P < 0.001) had a significantly higher probability of outpatient utilization, whereas those with middle school education (OR = 0.61, P < 0.001) and better household economy (OR = 0.96, P < 0.001) had a significantly less likelihood of using outpatient care. Rural patients (β = -0.72, P < 0.05) may have lower OOPE, while those with better household economy (β = 0.29, P < 0.05; β = 0.58, P < 0.05) and poor self-rated health (β = 0.62, P < 0.001) occurred higher OOPE. Regarding primary inpatient care, adults who were living in rural areas (OR = 1.48, P < 0.001), covered by Urban Employee Basic Medical Insurance (UEBMI) or Urban Rural Basic Medical Insurance (URBMI) (OR = 2.46, P < 0.001; OR = 1.81, P < 0.001) and with poor self-rated health (OR = 2.30, P < 0.001) had a significantly higher probability of using inpatient care, whereas individuals who were female (OR = 0.74, P < 0.001), with middle school education (OR = 0.40, P < 0.001) and better household economy (OR = 0.04, P < 0.001) had a significantly lower tendency to use inpatient care. Significantly, more OOPE occurred by individuals who were women (β = 0.18, P < 0.05) and with better household economy (β = 0.40, P < 0.001; β = 0.62, P < 0.001), whereas those who were covered by URBMI (β = -0.25, P < 0.05) and satisfied with their health (β = -0.21, P < 0.05) had less OOPE.

Conclusion

To prompt primary care visits and reduce economic burden among subgroups, more policy support is in need, such as tilting professional medical staff and funding to rural areas, enhancing awareness of disease prevention among vulnerable groups and so on.

Factors associated with formal and informal resource utilization in nursing home patients with and without dementia: cross-sectional analyses from the COSMOS trial

OBJECTIVES

To investigate the association between clinical, demographic, and organizational factors and formal (health professionals) and informal (relatives) resource utilization in nursing home patients with and without dementia.

METHODS

Baseline data from the multicomponent cluster randomized control COSMOS trial including 33 Norwegian nursing homes and 723 residents with and without dementia. Nursing home staff (n = 117) participated as proxy raters to approximate formal and informal resource use in daily care.

MEASUREMENTS

The primary outcome was the Resource Utilization in Dementia - Formal Care scale to assess formal and informal care time in hours/month regarding basic activities of daily living (ADL), instrumental ADL, and supervision. Secondary outcomes were hours/week spent on formal and informal leisure activities. Behavioral and psychological symptoms in dementia (BPSD) were assessed by the Neuropsychiatric Inventory-Nursing Home version, physical function by the Physical Self-Maintenance Scale, and psychotropic drug use by the Anatomical Therapeutic Chemical classification system. Organizational factors were ward size and staff ratio.

RESULTS

Generalized linear mixed-effect models and two-part modelling revealed an association between increased formal care time and poorer physical function, higher agitation and psychotropic drug use and lower cognitive function (all p < .05). Enhanced formal leisure time was related to better ADL function (p < .05) and smaller wards (p < .05). The family related leisure time was associated with agitation, decline in ADL function, smaller wards, and better staffing ratio (all p < .05). Married patients received more informal direct care (p < .05) and leisure time (p < .05) compared to unmarried/widowed.

CONCLUSION

For nursing home staff, higher agitation and psychotropic drug use, and lower cognitive function, is associated with more direct care time, whereas leisure time activities are less prioritized in people with lower physical function. Informal caregivers' engagement is encouraged by smaller nursing homes and better staff ratio. Therefore, we recommend stakeholders and healthcare professionals to consider these clinical and organizational factors to optimize treatment and leisure time activities in nursing home patients with various needs.

TRIAL REGISTRATION

ClinicalTrials.gov ; NCT02238652.

Count data models for outpatient health services utilisation

BACKGROUND

Count data from the national survey captures healthcare utilisation within a specific reference period, resulting in excess zeros and skewed positive tails. Often, it is modelled using count data models. This study aims to identify the best-fitting model for outpatient healthcare utilisation using data from the Malaysian National Health and Morbidity Survey 2019 (NHMS 2019) and utilisation factors among adults in Malaysia.

METHODS

The frequency of outpatient visits is the dependent variable, and instrumental variable selection is based on Andersen's model. Six different models were used: ordinary least squares (OLS), Poisson regression, negative binomial regression (NB), inflated models: zero-inflated Poisson, marginalized-zero-inflated negative binomial (MZINB), and hurdle model. Identification of the best-fitting model was based on model selection criteria, goodness-of-fit and statistical test of the factors associated with outpatient visits.

RESULTS

The frequency of zero was 90%. Of the sample, 8.35% of adults utilized healthcare services only once, and 1.04% utilized them twice. The mean-variance value varied between 0.14 and 0.39. Across six models, the zero-inflated model (ZIM) possesses the smallest log-likelihood, Akaike information criterion, Bayesian information criterion, and a positive Vuong corrected value. Fourteen instrumental variables, five predisposing factors, six enablers, and three need factors were identified. Data overdispersion is characterized by excess zeros, a large mean to variance value, and skewed positive tails. We assumed frequency and true zeros throughout the study reference period. ZIM is the best-fitting model based on the model selection criteria, smallest Root Mean Square Error (RMSE) and higher R2. Both Vuong corrected and uncorrected values with different Stata commands yielded positive values with small differences.

CONCLUSION

State as a place of residence, ethnicity, household income quintile, and health needs were significantly associated with healthcare utilisation. Our findings suggest using ZIM over traditional OLS. This study encourages the use of this count data model as it has a better fit, is easy to interpret, and has appropriate assumptions based on the survey methodology.

Symptom Presence and Symptom Severity as Unique Indicators of Psychopathology: An Application of Multidimensional Zero-Inflated and Hurdle Graded Response Models

Questionnaires inquiring about psychopathology symptoms often produce data with excess zeros or the equivalent (e.g., none, never, and not at all). This type of zero inflation is especially common in nonclinical samples in which many people do not exhibit psychopathology, and if unaccounted for, can result in biased parameter estimates when fitting latent variable models. In the present research, we adopt a maximum likelihood approach in fitting multidimensional zero-inflated and hurdle graded response models to data from a psychological distress measure. These models include two latent variables: susceptibility, which relates to the probability of endorsing the symptom at all, and severity, which relates to the frequency of the symptom, given its presence. After estimating model parameters, we compute susceptibility and severity scale scores and include them as explanatory variables in modeling health-related criterion measures (e.g., suicide attempts, diagnosis of major depressive disorder). Results indicate that susceptibility and severity uniquely and differentially predict other health outcomes, which suggests that symptom presence and symptom severity are unique indicators of psychopathology and both may be clinically useful. Psychometric and clinical implications are discussed, including scale score reliability.

Long-Term Care Home Size Association with COVID-19 Infection and Mortality in Catalonia in March and April 2020

We aim to assess how COVID-19 infection and mortality varied according to facility size in 965 long-term care homes (LTCHs) in Catalonia during March and April 2020. We measured LTCH size by the number of authorised beds. Outcomes were COVID-19 infection (at least one COVID-19 case in an LTCH) and COVID-19 mortality. Risks of these were estimated with logistic regression and hurdle models. Models were adjusted for county COVID-19 incidence and population, and LTCH types. Sixty-five per cent of the LTCHs were infected by COVID-19. We found a strong association between COVID-19 infection and LTCH size in the adjusted analysis (from 45% in 10-bed homes to 97.5% in those with over 150 places). The average COVID-19 mortality in all LTCHs was 6.8% (3887 deaths) and 9.2% among the COVID-19-infected LTCHs. Very small and large homes had higher COVID-19 mortality, whereas LTCHs with 30 to 70 places had the lowest level. COVID-19 mortality sharply increased with LTCH size in counties with a cumulative incidence of COVID-19 which was higher than 250/100,000, except for very small homes, but slightly decreased with LTCH size when the cumulative incidence of COVID-19 was lower. To prevent infection and preserve life, the optimal size of an LTCH should be between 30 and 70 places.

An evaluation of the outcome metric 'days alive and at home' in older patients after hip fracture surgery

'Days alive and at home' is a validated measure that estimates the time spent at home, defined as the place of residence before admission to hospital. We evaluated this metric in older adults after hip fracture surgery and assessed two follow-up durations, 30 and 90 days. Patients aged ≥ 70 years who underwent hip fracture surgery were identified retrospectively via hospital admission and government mortality records. Patients who successfully returned home and were still alive within 90 days of surgery were distinguished from those who were not. Regression models were used to examine which variables were associated with failure to return home and number of days at home among those who did return, within 90 days of surgery. We analysed the records of 825 patients. Median (IQR [range]) number of days at home within 90 days (n = 788) was 54 (0-76 [0-88]) days and within 30 days (n = 797) it was 2 (0-21 [0-28]) days. Out of these, 274 (35%) patients did not return home within 90 days and 374 (47%) within 30 days after surgery. Known peri-operative risk-factors such as older age, pre-operative anaemia and postoperative acute renal impairment were associated with failure to return home. This study supports days alive and at home as a useful patient-centred outcome measure in older adults after hip fracture surgery. We recommend that this metric should be used in clinical trials and measured at 90, rather than 30, postoperative days. As nearly half of this patient population did not return home within 30 days, the shorter time-period catches fewer meaningful events.

The relationship between moderate to vigorous physical activity and metabolic syndrome: a Bayesian measurement error approach

Metabolic Syndrome (MetS) is a serious condition that can be an early warning sign of heart disease and Type 2 diabetes. MetS is characterized by having elevated levels of blood pressure, cholesterol, waist circumference, and fasting glucose. There are many articles in the literature exploring the relationship between physical activity and MetS, but most do not consider the measurement error in the physical activity measurements nor the correlations among the MetS risk factors. Furthermore, previous work has generally treated MetS as binary, rather than directly modeling the risk factors on their measured, continuous space. Using data from the National Health and Nutrition Examination Survey (NHANES), we explore the relationship between minutes of moderate to vigorous physical activity (MVPA) and MetS risk factors. We construct a measurement error model for the accelerometry data, and then model its relationship between MetS risk factors with nonlinear seemingly unrelated regressions, incorporating dependence among MetS risk factors. The novel features of this model give the medical research community a new way to understand relationships between MVPA and MetS. The results of this approach present the field with a different modeling perspective than previously taken and suggest future avenues of scientific discovery.

Comparison of Ambulatory Health Care Costs and Use Associated With Roux-en-Y Gastric Bypass vs Sleeve Gastrectomy

IMPORTANCE

Studies comparing contemporary bariatric surgical types could facilitate procedure selection for patients interested in reducing their frequency of health care visits and reliance on prescription drugs.

OBJECTIVE

To compare the association of sleeve gastrectomy (SG) and Roux-en-Y gastric bypass (RYGB) with ambulatory health care costs and use for as long as 4 years after surgery.

DESIGN, SETTING, AND PARTICIPANTS

This comparative effectiveness study, which included patients undergoing bariatric surgery who were aged 18 to 64 years with at least 24 months of enrollment data before surgery and 12 months of enrollment data after surgery, used a retrospective interrupted time series with a comparison group. Data represent insurance claims dated January 2006 to June 2017, with analyses completed in September 2021. Data were collected from US commercial and Medicare Advantage claims database. Cohorts were matched on characteristics including baseline body mass index category, diabetes status, baseline ambulatory care costs, region of the United States, and year of surgery.

EXPOSURES

SG or RYGB, based on procedure codes.

MAIN OUTCOMES AND MEASURES

Annual ambulatory health care costs, and subtypes of cost and use including prescriptions, office visits, laboratory encounters, and radiology.

RESULTS

Matched cohorts included 3049 patients who underwent SG and 3251 patients who underwent RYGB, with a mean (SD) age of 45.2 (10.0) years; 4820 (77%) were women. Full follow-up was 37% for SG (514 patients) and 38% for RYGB (643 patients) among those eligible for 4-year follow-up. There were no significant differences between SG and RYGB in total ambulatory costs, office visit costs, or radiology costs in all follow-up years. Patients who underwent SG had significantly higher prescription costs than those who underwent RYGB bypass in year 4 ($852.8 per patient per year; 95% CI: $395.6-$1310.0 per patient per year) with more cardiometabolic medication fills in each year (eg, year 4: 42.5%; 95% CI, 13.7%-71.2%). In contrast, early after surgery, patients who underwent SG had relatively fewer specialist visits (eg, year 1: -7.2%; 95% CI, -14.3% to -0.2%) and lower laboratory costs (eg, year 1: -$118.9 per patient per year; 95% CI, -$220.2 to -$17.5 per patient per year).

CONCLUSIONS AND RELEVANCE

Despite clinical studies showing greater weight loss and comorbidity improvement with RYGB vs SG, this study found no difference in total ambulatory costs for as long as 4 years after SG and RYGB. These findings may reflect the trade-off between greater improvements in cardiometabolic health and additional surgery-related care among patients undergoing RYGB. Studies with longer follow-up time could determine whether greater sustained weight loss from RYGB eventually results in lower costs compared with SG.

Risk Factors of Skilled Nursing Facility Admissions and the Interrelation With Hospitalization and Amount of Informal Caregiving Received

BACKGROUND

The correlations between skilled nursing facility (SNF) admissions, number of hospitalizations, and informal caregiving hours received after adjusting for physical and cognitive function and sociodemographic covariates are not well understood.

OBJECTIVE

The objective of this study was to better understand risk factors for SNF admissions and the interrelation with hospitalizations and amount of informal caregiving received, this study applied a novel joint modeling analysis to simultaneously explore the correlation and shared information between the 3 outcomes.

RESEARCH DESIGN

This was an observational follow-up study.

SUBJECTS

Data from 4836 older Americans included in the 2011-2015 rounds of the National Health and Aging Trends Study were linked with Centers for Medicare & Medicaid Services.

MEASURES

We jointly modeled SNF admission, hospital admissions, and informal caregiving hours received while accounting for possible risk factors. We addressed missing values by multiple imputation with chained equations.

RESULTS

SNF admission evidenced a strong positive correlation with hospital admission, and SNF admission evidenced a weak positive correlation with the informal caregiving hours received after adjustment for important risk factors. Non-Hispanic White race/ethnicity, living alone, not being Medicaid eligible, Alzheimer disease and related dementias diagnosis, activities of daily living disabilities, and frailty were associated with increased risk of SNF admissions and any/number of hospital admission. Lower educational level was also associated with the latter. Medicaid eligibility was the only factor not associated with any nor numbers of informal caregiving hours received.

CONCLUSIONS

Sociodemographic and health factors were important for predicting SNF admissions. After adjustment for important risk factors, SNF evidenced a strong positive correlation with the number of hospitalizations and a weak positive correlation with the hours of informal caregiving received.

On testing for homogeneity with zero-inflated models through the lens of model misspecification

In many applications of two-component mixture models such as the popular zero-inflated model for discrete-valued data, it is customary for the data analyst to evaluate the inherent heterogeneity in view of observed data. To this end, the score test, acclaimed for its simplicity, is routinely performed. It has long been recognized that this test may behave erratically under model misspecification, but the implications of this behavior remain poorly understood for popular two-component mixture models. For the special case of zero-inflated count models, we use data simulations and theoretical arguments to evaluate this behavior and discuss its implications in settings where the working model is restrictive with regard to the true data generating mechanism. We enrich this discussion with an analysis of count data in HIV research, where a one-component model is shown to fit the data reasonably well despite apparent extra zeros. These results suggest that a rejection of homogeneity does not imply that the underlying mixture model is appropriate. Rather, such a rejection simply implies that the mixture model should be carefully interpreted in the light of potential model misspecifications, and further evaluated against other competing models.

The Coordination Toolkit and Coaching Project: Cluster-Randomized Quality Improvement Initiative to Improve Patient Experience of Care Coordination

BACKGROUND

Given persistent gaps in coordination of care for medically complex primary care patients, efficient strategies are needed to promote better care coordination.

OBJECTIVE

The Coordination Toolkit and Coaching project compared two toolkit-based strategies of differing intensity to improve care coordination at VA primary care clinics.

DESIGN

Multi-site, cluster-randomized QI initiative.

PARTICIPANTS

Twelve VA primary care clinics matched in 6 pairs.

INTERVENTIONS

We used a computer-generated allocation sequence to randomize clinics within each pair to two implementation strategies. Active control clinics received an online toolkit with evidence-based tools and QI coaching manual. Intervention clinics received the online toolkit plus weekly assistance from a distance coach for 12 months.

MAIN MEASURES

We quantified patient experience of general care coordination using the Health Care System Hassles Scale (primary outcome) mailed at baseline and 12-month follow-up to serial cross-sectional patient samples. We measured the difference-in-difference (DiD) in clinic-level-predicted mean counts of hassles between coached and non-coached clinics, adjusting for clustering and patient characteristics using zero-inflated negative binomial regression and bootstrapping to obtain 95% confidence intervals. Other measures included care coordination QI projects attempted, tools adopted, and patient-reported exposure to projects.

KEY RESULTS

N = 2,484 (49%) patients completed baseline surveys and 2,481 (48%) completed follow-ups. Six coached clinics versus five non-coached clinics attempted QI projects. All coached clinics versus two non-coached clinics attempted more than one project or projects that were multifaceted (i.e., involving multiple components addressing a common goal). Five coached versus three non-coached clinics used 1-2 toolkit tools. Both the coached and non-coached clinics experienced pre-post reductions in hassle counts over the study period (- 0.42 (- 0.76, - 0.08) non-coached; - 0.40 (- 0.75, - 0.06) coached). However, the DiD (0.02 (- 0.47, 0.50)) was not statistically significant; coaching did not improve patient experience of care coordination relative to the toolkit alone.

CONCLUSION

Although coached clinics attempted more or more complex QI projects and used more tools than non-coached clinics, coaching provided no additional benefit versus the online toolkit alone in patient-reported outcomes.

TRIAL REGISTRATION

ClinicalTrials.gov identifier: NCT03063294.

Bayesian spatio-temporal distributed lag modeling for delayed climatic effects on sparse malaria incidence data

Background

In many areas of the Greater Mekong Subregion (GMS), malaria endemic regions have shrunk to patches of predominantly low-transmission. With a regional goal of elimination by 2030, it is important to use appropriate methods to analyze and predict trends in incidence in these remaining transmission foci to inform planning efforts. Climatic variables have been associated with malaria incidence to varying degrees across the globe but the relationship is less clear in the GMS and standard methodologies may not be appropriate to account for the lag between climate and incidence and for locations with low numbers of cases.

Methods

In this study, a methodology was developed to estimate the spatio-temporal lag effect of climatic factors on malaria incidence in Thailand within a Bayesian framework. A simulation was conducted based on ground truth of lagged effect curves representing the delayed relation with sparse malaria cases as seen in our study population. A case study to estimate the delayed effect of environmental variables was used with malaria incidence at a fine geographic scale of sub-districts in a western province of Thailand.

Results

From the simulation study, the model assumptions which accommodated both delayed effects and excessive zeros appeared to have the best overall performance across evaluation metrics and scenarios. The case study demonstrated lagged climatic effect estimation of the proposed modeling with real data. The models appeared to be useful to estimate the shape of association with malaria incidence.

Conclusions

A new method to estimate the spatiotemporal effect of climate on malaria trends in low transmission settings is presented. The developed methodology has potential to improve understanding and estimation of past and future trends in malaria incidence. With further development, this could assist policy makers with decisions on how to more effectively distribute resources and plan strategies for malaria elimination.

Studying dietary intake in daily life through multilevel two-part modelling: a novel analytical approach and its practical application

BACKGROUND

Understanding which factors influence dietary intake, particularly in daily life, is crucial given the impact diet has on physical as well as mental health. However, a factor might influence whether but not how much an individual eats and vice versa or a factor's importance may differ across these two facets. Distinguishing between these two facets, hence, studying dietary intake as a dual process is conceptually promising and not only allows further insights, but also solves a statistical issue. When assessing the association between a predictor (e.g. momentary affect) and subsequent dietary intake in daily life through ecological momentary assessment (EMA), the outcome variable (e.g. energy intake within a predefined time-interval) is semicontinuous. That is, one part is equal to zero (i.e. no dietary intake occurred) and the other contains right-skewed positive values (i.e. dietary intake occurred, but often only small amounts are consumed). However, linear multilevel modelling which is commonly used for EMA data to account for repeated measures within individuals cannot be applied to semicontinuous outcomes. A highly informative statistical approach for semicontinuous outcomes is multilevel two-part modelling which treats the outcome as generated by a dual process, combining a multilevel logistic/probit regression for zeros and a multilevel (generalized) linear regression for nonzero values.

METHODS

A multilevel two-part model combining a multilevel logistic regression to predict whether an individual eats and a multilevel gamma regression to predict how much is eaten, if an individual eats, is proposed. Its general implementation in R, a widely used and freely available statistical software, using the R-package brms is described. To illustrate its practical application, the analytical approach is applied exemplary to data from the Eat2beNICE-APPetite-study.

RESULTS

Results highlight that the proposed multilevel two-part model reveals process-specific associations which cannot be detected through traditional multilevel modelling.

CONCLUSIONS

This paper is the first to introduce multilevel two-part modelling as a novel analytical approach to study dietary intake in daily life. Studying dietary intake through multilevel two-part modelling is conceptually as well as methodologically promising. Findings can be translated to tailored nutritional interventions targeting either the occurrence or the amount of dietary intake.

Appropriate analyses of bimodal substance use frequency outcomes: a mixture model approach

Background: In addiction research, outcome measures are often characterized by bimodal distributions. One mode can be for individuals with low substance use and the other mode for individuals with high substance use. Applying standard statistical procedures to bimodal data may result in invalid inference. Mixture models are appropriate for bimodal data because they assume that the sampled population is composed of several underlying subpopulations.Objectives: To introduce a novel mixture modeling approach to analyze bimodal substance use frequency data.Methods: We reviewed existing models used to analyze substance use frequency outcomes and developed multiple alternative variants of a finite mixture model. We applied all methods to data from a randomized controlled study in which 30-day alcohol abstinence was the primary outcome. Study data included 73 individuals (38 men and 35 women). Models were implemented in the software packages SAS, Stata, and Stan.Results: Shortcomings of existing approaches include: 1) inability to model outcomes with multiple modes, 2) invalid statistical inferences, including anti-conservative p-values, 3) sensitivity of results to the arbitrary choice to model days of substance use versus days of substance abstention, and 4) generation of predictions outside the range of common substance use frequency outcomes. Our mixture model variants avoided all of these shortcomings.Conclusions: Standard models of substance use frequency outcomes can be problematic, sometimes overstating treatment effectiveness. The mixture models developed improve the analysis of bimodal substance use frequency.

Hospital-free Days: A Pragmatic and Patient-centered Outcome for Trials Among Critically and Seriously Ill Patients

Hospital-free days (HFDs), alternatively known as "days alive and outside the hospital," is increasingly used as a primary or secondary outcome in randomized trials among critically and seriously ill patients. This novel outcome measure addresses an existing gap in the availability of patient-centered, reliably obtained outcome measures among patients with acute respiratory failure, advanced lung diseases, lung transplantation, and other serious and critical illnesses. Traditional outcomes such as mortality, organ-failure-free days, and longitudinal patient-reported measures have distinct drawbacks that limit their suitability as endpoints in trials of patients with serious illness, particularly those trials with pragmatic designs. By contrast, HFDs provides a summary measure of important health events and is easily calculated from administrative or electronic health record data, thereby balancing the goals of patient-centeredness and pragmatic measurement. However, before HFDs can be widely adopted as an endpoint in trials of patients with respiratory and critical illnesses, several questions must be addressed regarding the optimal definition, measurement, and analysis of HFDs. In this perspective, we outline important considerations relevant to the use of HFDs as a trial endpoint and suggest directions for further development of the measure.

A comparison of zero-inflated and hurdle models for modeling zero-inflated count data

Counts data with excessive zeros are frequently encountered in practice. For example, the number of health services visits often includes many zeros representing the patients with no utilization during a follow-up time. A common feature of this type of data is that the count measure tends to have excessive zero beyond a common count distribution can accommodate, such as Poisson or negative binomial. Zero-inflated or hurdle models are often used to fit such data. Despite the increasing popularity of ZI and hurdle models, there is still a lack of investigation of the fundamental differences between these two types of models. In this article, we reviewed the zero-inflated and hurdle models and highlighted their differences in terms of their data generating processes. We also conducted simulation studies to evaluate the performances of both types of models. The final choice of regression model should be made after a careful assessment of goodness of fit and should be tailored to a particular data in question.

Implantable Cardioverter-Defibrillator Shocks During COVID-19 Outbreak

Background

COVID‐19 was temporally associated with an increase in out‐of‐hospital cardiac arrests, but the underlying mechanisms are unclear. We sought to determine if patients with implantable defibrillators residing in areas with high COVID‐19 activity experienced an increase in defibrillator shocks during the COVID‐19 outbreak.

Methods and Results

Using the Medtronic (Mounds View, MN) Carelink database from 2019 and 2020, we retrospectively determined the incidence of implantable defibrillator shock episodes among patients residing in New York City, New Orleans, LA, and Boston, MA. A total of 14 665 patients with a Medtronic implantable defibrillator (age, 66±13 years; and 72% men) were included in the analysis. Comparing analysis time periods coinciding with the COVID‐19 outbreak in 2020 with the same periods in 2019, we observed a larger mean rate of defibrillator shock episodes per 1000 patients in New York City (17.8 versus 11.7, respectively), New Orleans (26.4 versus 13.5, respectively), and Boston (30.9 versus 20.6, respectively) during the COVID‐19 surge. Age‐ and sex‐adjusted hurdle model showed that the Poisson distribution rate of defibrillator shocks for patients with ≥1 shock was 3.11 times larger (95% CI, 1.08–8.99; P=0.036) in New York City, 3.74 times larger (95% CI, 0.88–15.89; P=0.074) in New Orleans, and 1.97 times larger (95% CI, 0.69–5.61; P=0.202) in Boston in 2020 versus 2019. However, the binomial odds of any given patient having a shock episode was not different in 2020 versus 2019.

Conclusions

Defibrillator shock episodes increased during the higher COVID‐19 activity in New York City, New Orleans, and Boston. These observations may provide insights into COVID‐19–related increase in cardiac arrests.

Two-part quantile regression models for semi-continuous longitudinal data: A finite mixture approach

This article develops a two-part finite mixture quantile regression model for semi-continuous longitudinal data. The proposed methodology allows heterogeneity sources that influence the model for the binary response variable to also influence the distribution of the positive outcomes. As is common in the quantile regression literature, estimation and inference on the model parameters are based on the asymmetric Laplace distribution. Maximum likelihood estimates are obtained through the EM algorithm without parametric assumptions on the random effects distribution. In addition, a penalized version of the EM algorithm is presented to tackle the problem of variable selection. The proposed statistical method is applied to the well-known RAND Health Insurance Experiment dataset which gives further insights on its empirical behaviour.

Association between processes of diabetes care and health care utilization in patients with diabetes: Evidence from a nationally representative US sample

BACKGROUND

This study examined the association between quality of processes of diabetes care in terms of adherence to American Diabetes Association (ADA)-recommended guidelines and health care utilization in patients with diabetes.

METHODS

Adults with diabetes were identified from the pool of five panels of the Medical Expenditure Panel Survey, a nationally representative US sample, between 2012 and 2017. The Diabetes Care Survey was used to determine adherence to the ADA-recommended guidelines for processes of diabetes care if all of the following were performed annually: glycosylated hemoglobin check, foot examination, dilated eye examination, lipid panel, influenza immunization, blood pressure check, and dental examination. Health care utilization in terms of inpatient hospitalization, and emergency department (ED) and outpatient visits were estimated using two-part hurdle models.

RESULTS

An estimated 26.3 million adults with diabetes were derived from the pooled 5-panel data, of which 7.8% met the ADA-recommended guidelines for processes of diabetes care, and adherence rates of individual recommendations were generally below 50%. Overall, adults who adhered to the ADA-recommendations were older, non-Hispanic white, and married nonsmokers with private insurance and higher income. Mean inpatient hospital stays, ED, and outpatient visits between ADA-adherent vs nonadherent patients were 0.98 vs 1.62 (P < .001), 0.36 vs 0.39 (P = .074), and 17.9 vs 12.8 (P < .001), respectively.

CONCLUSIONS

Socioeconomic disadvantage and minority status were linked with nonadherence to the ADA-recommended processes of diabetes care. Adherence to the ADA recommendation was associated with significant reduction in inpatient hospitalization and a trend toward less ED visits. Our findings may apply to the United States and are likely to be different in other parts of the world.

Analysis of the Railway Accident-Related Damages in South Korea

Railway accidents are critical issues characterized by a large number of injuries and fatalities per accident due to massive public transport systems. This study proposes a new approach for evaluating the damages resulting from railway accidents using the two-part models (TPMs) such as the zero-inflated Poisson regression model (ZIP model) and the zero-inflated negative-binomial regression model (ZINB model) for the non-negative count measurements and the zero-inflated gamma regression model (ZIG model) and the zero-inflated log-normal regression model (ZILN model) for the semi-continuous measurements. The models are employed for the evaluation of the railway accidents on Korea Railroad, considering the accident damages, such as the train delay time, the number of trains delayed and the cost of considering the accident count responses, for the period 2008 to 2016. From the results obtained, we found that the human-related factors, the high-speed railway system or the Korea Train Express (KTX) and the number of casualties, are the main cost-escalating factors. The number of trains delayed and the amount of delay time tend to increase both the probability of incurring costs and the amount of cost. For better evaluation, the railway accident data should contain accurate information with less recurrence of zeros.

Opioid Prescribing in the 2016 Medicare Fee-for-Service Population

BACKGROUND AND OBJECTIVES

Opioid use and misuse are prevalent and remain a national crisis. This study identified beneficiary characteristics associated with filling opioid prescriptions, variation in opioid dosing, and opioid use with average daily doses (ADDs) equal to 120 morphine milligram equivalents (MMEs) or more in the 100% Medicare fee-for-service (FFS) population.

DESIGN, SETTING, PARTICIPANTS, AND MEASUREMENTS

In a cohort of FFS beneficiaries with 12 months of Medicare Part D coverage in 2016, we examined patient factors associated with filling an opioid prescription (n = 20,880,490) and variation in ADDs (n = 7,325,031) in a two-part model. Among those filling opioids, we also examined the probability of ADD equal to 120 MMEs or more via logistic regression.

RESULTS

About 35% of FFS beneficiaries had one or more opioid prescription fills in 2016 and 1.5% had ADDs equal to 120 MMEs or more. Disability-eligible beneficiaries and beneficiaries with multiple chronic conditions were more likely to fill opioids, to have higher ADDs or were more likely to have ADD equal to 120 MMEs or more. Beneficiaries with chronic obstructive pulmonary disease (COPD) were more likely to fill opioids (odds ratio (OR) = 1.47, 95% confidence interval (CI) = 1.46-1.47), have higher ADDs (rate ratio = 1.06, 95% CI = 1.06-1.06) when filled and were more likely to have ADD equal to 120 MMEs or more (OR = 1.23, 95% CI = 1.21-1.24). Finally, black and Hispanic beneficiaries were less likely to fill opioids, had lower overall doses and were less likely to have ADDs equal to 120 MMEs or more compared to white beneficiaries.

CONCLUSION

Several beneficiary subgroups have underappreciated risk of adverse events associated with ADD equal to 120 MMEs or more that may benefit from opioid optimization interventions that balance pain management and adverse event risk, especially beneficiaries with COPD who are at risk for respiratory depression.

A new mixed-effects regression model for the analysis of zero-modified hierarchical count data

Count data sets are traditionally analyzed using the ordinary Poisson distribution. However, such a model has its applicability limited as it can be somewhat restrictive to handle specific data structures. In this case, it arises the need for obtaining alternative models that accommodate, for example, (a) zero-modification (inflation or deflation at the frequency of zeros), (b) overdispersion, and (c) individual heterogeneity arising from clustering or repeated (correlated) measurements made on the same subject. Cases (a)-(b) and (b)-(c) are often treated together in the statistical literature with several practical applications, but models supporting all at once are less common. Hence, this paper's primary goal was to jointly address these issues by deriving a mixed-effects regression model based on the hurdle version of the Poisson-Lindley distribution. In this framework, the zero-modification is incorporated by assuming that a binary probability model determines which outcomes are zero-valued, and a zero-truncated process is responsible for generating positive observations. Approximate posterior inferences for the model parameters were obtained from a fully Bayesian approach based on the Adaptive Metropolis algorithm. Intensive Monte Carlo simulation studies were performed to assess the empirical properties of the Bayesian estimators. The proposed model was considered for the analysis of a real data set, and its competitiveness regarding some well-established mixed-effects models for count data was evaluated. A sensitivity analysis to detect observations that may impact parameter estimates was performed based on standard divergence measures. The Bayesian p -value and the randomized quantile residuals were considered for model diagnostics.

Best (but oft-forgotten) practices: sample size and power calculation for a dietary intervention trial with episodically consumed foods

Dietary interventions often target foods that are underconsumed relative to dietary guidelines, such as vegetables, fruits, and whole grains. Because these foods are only consumed episodically for some participants, data from such a study often contains a disproportionally large number of zeros due to study participants who do not consume any of the target foods on the days that dietary intake is assessed, thus generating semicontinuous data. These zeros need to be properly accounted for when calculating sample sizes to ensure that the study is adequately powered to detect a meaningful intervention effect size. Nonetheless, this issue has not been well addressed in the literature. Instead, methods that are common for continuous outcomes are typically used to compute the sample sizes, resulting in a substantially under- or overpowered study. We propose proper approaches to calculating the sample size needed for dietary intervention studies that target episodically consumed foods. Sample size formulae are derived for detecting the mean difference in the amount of intake of an episodically consumed food between an intervention and a control group. Numerical studies are conducted to investigate the accuracy of the sample size formulae as compared with the ad hoc methods. The simulation results show that the proposed formulae are appropriate for estimating the sample sizes needed to achieve the desired power for the study. The proposed method for sample size is recommended for designing dietary intervention studies targeting episodically consumed foods.

A comparison of residual diagnosis tools for diagnosing regression models for count data

Background

Examining residuals is a crucial step in statistical analysis to identify the discrepancies between models and data, and assess the overall model goodness-of-fit. In diagnosing normal linear regression models, both Pearson and deviance residuals are often used, which are equivalently and approximately standard normally distributed when the model fits the data adequately. However, when the response vari*able is discrete, these residuals are distributed far from normality and have nearly parallel curves according to the distinct discrete response values, imposing great challenges for visual inspection.

Methods

Randomized quantile residuals (RQRs) were proposed in the literature by Dunn and Smyth (1996) to circumvent the problems in traditional residuals. However, this approach has not gained popularity partly due to the lack of investigation of its performance for count regression including zero-inflated models through simulation studies. Therefore, we assessed the normality of the RQRs and compared their performance with traditional residuals for diagnosing count regression models through a series of simulation studies. A real data analysis in health care utilization study for modeling the number of repeated emergency department visits was also presented.

Results

Our results of the simulation studies demonstrated that RQRs have low type I error and great statistical power in comparisons to other residuals for detecting many forms of model misspecification for count regression models (non-linearity in covariate effect, over-dispersion, and zero inflation). Our real data analysis also showed that RQRs are effective in detecting misspecified distributional assumptions for count regression models.

Conclusions

Our results for evaluating RQRs in comparison with traditional residuals provide further evidence on its advantages for diagnosing count regression models.

The Impact of Health Care Education on Utilization Among Adolescents Preparing for Emancipation From Foster Care

PURPOSE

As teens in foster care prepare for emancipation, health care navigation is often overlooked, as caseworkers address other social needs. This study examined the impact of health care education materials designed for foster youth, called ICare2CHECK. It was hypothesized that ICare2CHECK would increase nonurgent ambulatory health care use and decrease emergency/urgent care use.

METHODS

Adolescents (N = 151; aged 16-22 years) were enrolled in ICare2CHECK and received health education materials at their baseline study visit. Surveys were repeated every 3 months to assess health care utilization. After 12 months of enrollment, health care data for all eligible youth and matched comparison youth (N = 151) over the previous 24 months were extracted from the electronic health record (N = 302). Electronic health record data were coded as counts of completed nonurgent ambulatory care encounters (i.e., primary and preventative care and specialty care), completed urgent or emergency encounters (i.e., urgent and emergency department visits and hospitalizations), completed foster care clinic visits, and total completed visits.

RESULTS

Health care use significantly decreased over time for both enrolled and comparison youth. Females, youth engaging in health risk behaviors, and those with a mental health or chronic condition diagnosis used significantly more health care. Receipt of educational materials was associated with a smaller decline in health care use and nonurgent ambulatory care use, controlling for covariates. Self-reported use of educational materials was associated with increased utilization in the enrolled condition.

CONCLUSIONS

Results suggest that ICare2CHECK is associated with increased engagement in health care generally and nonurgent ambulatory care specifically (e.g., outpatient primary and specialty care).

Emergency attendance for acute hyper- and hypoglycaemia in the adult diabetic population of the metropolitan area of Milan: quantifying the phenomenon and studying its predictors

BACKGROUND

We quantified, among diabetic adults, the frequency, costs, and factors associated with visits to the emergency department (EDs) and subsequent hospitalizations for acute hypoglycaemic and hyperglycaemic events.

METHODS

We included adults with diabetes residing in the Milan Agency for Health Protection between 2015 and 2017. From healthcare databases, we identified demographic variables, comorbidities, type of treatment, insulin treatment duration, previous ED attendances for acute glycaemic events, and two indicators of glycaemic monitoring. Using a validated ICD-9-CM coding algorithm, we identified all ED attendances for acute glycaemic events from the ED database and calculated their incidence. We computed the direct costs from health databases and presented them as average annual mean costs for those having had at least an ED attendance. The analysis of the association between the number of ED attendances and potential determinants was performed using zero-inflated negative binomial regression models. These two-part models concomitantly estimate two sets of parameters: the odds-ratios (ORs) of having no attendances and the incidence rate ratios (IRRs) of attendance.

RESULTS

The cohort included 168,285 subjects, 70% of subjects were older than 64 years, 56% were males, and 26% were treated with insulin. The incidence of acute glycaemic events for those attending the ED was 7.0 per 1000 patient-years, followed by hospitalization 26.0% of the time. The total annual direct cost for ED attendances due to acute glycaemic events was 174,000 €. Type of antidiabetic treatment had the strongest association with ED attendances for hypoglycaemia. Patients assuming insulin only had a lower probability of having no attendances (OR compared to those who assumed non-insulin antidiabetic drugs =0.01, 95% CI = 0.00-0.02). These patients also had the highest rate of hyperglycaemic episodes (IRR = 7.7, 95% CI = 5.1-11.7 for insulin only vs. non-insulin antidiabetic drugs). Subjects having had a previous episode of the same type leading to an ED visit had a higher rate of subsequent attendances (IRR for hypoglycaemia = 5.3, 95% CI = 3.9-7.3 and IRR for hyperglycaemia = 3.7, 95% CI = 1.3-10.2).

CONCLUSION

Insulin treatment and having had a prior acute glycaemic event leading to an ED visit were major predictors of ED attendance for hyper and hypoglycaemia in a population of adults with diabetes.

Patient experience of health care system hassles: Dual-system vs single-system users

OBJECTIVE

To compare health care system problems or "hassles" experienced by Veterans receiving VA health care only versus those receiving dual care from both VA and non-VA community providers.

DATA SOURCES

We collected survey data in 2017-2018 from 2444 randomly selected Veterans with four or more primary care visits in the prior year at one of 12 VA primary care clinics located in four geographically diverse regions of the United States.

STUDY DESIGN

We used baseline surveys from the Coordination Toolkit and Coaching quality improvement project to explore Veterans' experience of hassles (dependent variable), source of health care, self-rated physical and mental health, and sociodemographics.

DATA COLLECTION

Participants responded to mailed surveys by mail, telephone, or online.

PRINCIPAL FINDINGS

The number of reported hassles ranged from 0 to 16; 79 percent of Veterans reported experiencing one or more hassles. Controlling for sociodemographic characteristics and self-rated physical and mental health, zero-inflated negative binominal regression indicated that dual care users experienced more hassles than VA-only users (adjusted predicted average 5.5 [CI: 5.2, 5.8] vs 4.3 [CI: 4.1, 4.6] hassles [P < .0001]).

CONCLUSIONS

Anticipated increases in Veterans accessing community-based care may require new strategies to help VA primary care teams optimize care coordination for dual care users.

A model-based approach for clustering of multivariate semicontinuous data with application to dietary pattern analysis and intervention

Semicontinuous data, characterized by a sizable number of zeros and observations from a continuous distribution, are frequently encountered in health research concerning food consumptions, physical activities, medical and pharmacy claims expenditures, and many others. In analyzing such semicontinuous data, it is imperative that the excessive zeros be adequately accounted for to obtain unbiased and efficient inference. Although many methods have been proposed in the literature for the modeling and analysis of semicontinuous data, little attention has been given to clustering of semicontinuous data to identify important patterns that could be indicative of certain health outcomes or intervention effects. We propose a Bernoulli-normal mixture model for clustering of multivariate semicontinuous data and demonstrate its accuracy as compared to the well-known clustering method with the conventional normal mixture model. The proposed method is illustrated with data from a dietary intervention trial to promote healthy eating behavior among children with type 1 diabetes. In the trial, certain diabetes friendly foods (eg, total fruit, whole fruit, dark green and orange vegetables and legumes, whole grain) were only consumed by a proportion of study participants, yielding excessive zero values due to nonconsumption of the foods. Baseline foods consumptions data in the trial are used to explore preintervention dietary patterns among study participants. While the conventional normal mixture model approach fails to do so, the proposed Bernoulli-normal mixture model approach has shown to be able to identify a dietary profile that significantly differentiates the intervention effects from others, as measured by the popular healthy eating index at the end of the trial.

Right-Site Care Programme with a community-based family medicine clinic in Singapore: secondary data analysis of its impact on mortality and healthcare utilisation

OBJECTIVE

Stable patients with chronic conditions could be appropriately cared for at family medicine clinics (FMC) and discharged from hospital specialist outpatient clinics (SOCs). The Right-Site Care Programme with Frontier FMC emphasised care organised around patients in community rather than hospital-based providers, with one identifiable primary provider. This study evaluated impact of this programme on mortality and healthcare utilisation.

DESIGN

A retrospective study without randomisation using secondary data analysis of patients enrolled in the intervention matched 1:1 with unenrolled patients as controls.

SETTING

Programme was supported by the Ministry of Health in Singapore, a city-state nation in Southeast Asia with 5.6 million population.

PARTICIPANTS

Intervention group comprises patients enrolled from January to December 2014 (n=684) and control patients (n=684) with at least one SOC and no FMC attendance during same period.

INTERVENTIONS

Family physician in Frontier FMC managed patients in consultation with relevant specialist physicians or fully managed patients independently. Care teams in SOCs and FMC used a common electronic medical records system to facilitate care coordination and conducted regular multidisciplinary case conferences.

PRIMARY OUTCOME MEASURES

Deidentified linked healthcare administrative data for time period of January 2011 to December 2017 were extracted. Three-year postenrolment mortality rates and utilisation frequencies and charges for SOC, public primary care centres (polyclinic), emergency department attendances and emergency, non-day surgery inpatient and all-cause admissions were compared.

RESULTS

Intervention patients had lower mortality rate (HR=0.37, p<0.01). Among those with potential of postenrolment polyclinic attendance, intervention patients had lower frequencies (incidence rate ratio (IRR)=0.60, p<0.01) and charges (mean ratio (MR)=0.51, p<0.01). Among those with potential of postenrolment SOC attendance, intervention patients had higher frequencies (IRR=2.06, p<0.01) and charges (MR=1.86, p<0.01).

CONCLUSIONS

Intervention patients had better survival, probably because their chronic conditions were better managed with close monitoring, contributing to higher total outpatient attendance frequencies and charges.

Inequity in healthcare needs, health service use and financial burden of medical expenditures in China: results from a consecutive household monitoring study in Jiangsu Province

Background

Although public medical insurance covers over 95% of the population in China, disparities in health service use and out-of-pocket (OOP) health expenditure across income groups are still widely observed. This study aims to investigate the socio-economic disparities in perceived healthcare needs, informal care, formal care and payment for healthcare and explore their equity implication.

Methods

We assessed healthcare needs, service use and payment in 400 households in rural and urban areas in Jiangsu, China, and included only the adult sample (N = 925). One baseline survey and 10 follow-up surveys were conducted during the 7-month monitoring period, and the Affordability Ladder Program (ALP) framework was adopted for data analysis. Negative binomial/zero-inflated negative binomial and logit regression models were used to explore factors associated with perceived needs of care and with the use of self-treatment, outpatient and inpatient care respectively. Two-part model and logit regression modeling were conducted to explore factors associated with OOP health expenditure and with the likelihood of incurring catastrophic health expenditure (CHE).

Results

After adjusting for covariates, rural residence was significantly associated with more perceived healthcare needs, more self-treatment, higher probability of using outpatient and inpatient service, more OOP health expenditure and higher likelihood of incurring catastrophic expenditure (P < 0.05). Compared to the Urban Employee Basic Medical Insurance (UEBMI), enrollment in the New Rural Cooperative Medical Scheme (NRCMS) or in the Urban Resident Basic Medical Insurance (URBMI) was correlated with lower probability of ever using outpatient services, but with more outpatient visits when people were at risk of using outpatient service (P < 0.05). NRCMS/URBMI enrollment was also associated with higher likelihood of incurring CHE compared to UEBMI enrollment (OR = 2.02, P < 0.05); in stratified analysis of the rural and urban sample this effect was only significant for the rural population.

Conclusions

The rural population in Jiangsu perceived more healthcare needs, had a higher probability of using both informal and formal healthcare services, and had more OOP health expenditure and a higher likelihood of incurring CHE. The inequity mainly exists in health care financing, and may be partially addressed through improving the benefit packages of NRCMS/URBMI.

Level of Agreement Between Problem Gamblers' and Collaterals' Reports: A Bayesian Random-Effects Two-Part Model

This study investigates the level of agreement between problem gamblers and their concerned significant others (CSOs) regarding the amount of money lost when gambling. Reported losses were analyzed from 266 participants (133 dyads) seeking treatment, which included different types of CSO-gambler dyads. The intraclass correlation coefficients (ICCs) concerning the money lost when gambling during the last 30 days were calculated based on the timeline followback. In order to model reports that were highly skewed and included zeros, a two-part generalized linear mixed-effects model was used. The results were compared from models assuming either a Gaussian, two-part gamma, or two-part lognormal response distribution. Overall, the results indicated a fair level of agreement, ICC = .57, 95% CI (.48, .64), between the gamblers and their CSOs. The partner CSOs tended to exhibit better agreement than the parent CSOs with regard to the amount of money lost, ICCdiff = .20, 95% CI (.03, .39). The difference became smaller and inconclusive when reports of no losses (zeros) were included, ICCdiff = .16, 95% CI (- .05, .36). A small simulation investigation indicated that the two-part model worked well under assumptions related to this study, and further, that calculating the ICCs under normal assumptions led to incorrect conclusions regarding the level of agreement for skewed reports (such as gambling losses). For gambling losses, the normal assumption is unlikely to hold and ICCs based on this assumption are likely to be highly unreliable.

Zero-augmented accelerated spatial failure model for modeling hospital length of stay data

Hospital length of stay (LOS) is often used as an indicator for hospital efficiency and resource utilization. LOS is nonnegative with presence of zeros and typically positively skewed with a long right tail, which may not be adequately modelled by traditional distributions, such as lognormal. We developed a zero-augmented accelerated frailty model for modeling the extreme skewness with the presence of zeros. Levels of utilization of health services may vary geographically, so conditional autoregressive priors were used to provide spatial smoothing across neighboring hospital health districts. The random effect terms are further linked to investigate if the capacity for longer LOS are consistently higher or lower at the health district level. Modeling and inference used the Bayesian approach via Markov Chain Monte Carlo simulation techniques. We demonstrated the proposed model for modeling the LOS of patients admitted due to chronic lower respiratory disease in Saskatchewan, Canada.