A Multinomial Regression Approach to Model Outcome Heterogeneity

Primary Care Asthma Attack Prediction Models for Adults: A Systematic Review of Reported Methodologies and Outcomes

Prognostic models hold great potential for predicting asthma exacerbations, providing opportunities for early intervention, and are a popular area of current research. However, it is unclear how models should be compared and contrasted, given their differences in both design and performance, particularly with a view to potential implementation in routine practice. This systematic review aimed to identify novel predictive models of asthma attacks in adults and compare differences in construction related to populations, outcome definitions, prediction time horizons, algorithms, validation, and performance estimation. Twenty-five studies were identified for comparison, with varying definitions of asthma attacks and prediction event time horizons ranging from 15 days to 30 months. The most commonly used algorithm was logistic regression (20/25 studies); however, none of the six which tested multiple algorithms identified it as highest performing algorithm. The effect of various study design characteristics on performance was evaluated in order to provide context to the limitations of highly performing models. Models used a variety of constructs, which affected both their performance and their viability for implementation in routine practice. Consultation with stakeholders is necessary to identify priorities for model refinement and to create a benchmark of acceptable performance for implementation in clinical practice.

Development and Early Experience of a Primary Care Learning Collaborative in a Large Health Care System

Introduction:

Primary care clinicians are presented with hundreds of new clinical recommendations and guidelines. To consider practice change clinicians must identify relevant information and develop a contextual framework. Too much attention to information irrelevant to one’s practice results in wasted resources. Too little results in care gaps. A small group of primary care clinicians in a large health system sought to address the problem of vetting new information and providing peer reviewed context. This was done by engaging colleagues across the system though a primary care learning collaborative.

Methods:

The collaborative was a grass roots initiative between community and academic-based clinicians. They invited all the system’s primary care clinicians to participate. They selected new recommendations or guidelines and used surveys as the principal communication instrument. Surveys shared practice experience and also invited members to give narrative feedback regarding their acceptance of variation in care relate to the topic. A description of the collaborative along with its development, processes, and evolution are discussed. Process changes to address needs during the COVID-19 pandemic including expanded information sharing was necessary.

Results:

Collaborative membership reached across 5 states and included family medicine, internal medicine, and pediatrics. Members found involvement with the collaborative useful. Less variation in care was thought important for public health crises: the COVID pandemic and opioid epidemic. Greater practice variation was thought acceptable for adherence to multispecialty guidelines, such as diabetes, lipid management, and adult ADHD care. Process changes during the pandemic resulted in more communications between members to avoid practice gaps.

Conclusion:

An internet-based learning collaborative in a health system had good engagement from its members. Using novel methods, it was able to provide members with feedback related to the importance of new practice recommendations as perceived by their peers. Greater standardization was thought necessary when adopting measures to address public health crisis, and less necessary when addressing multispecialty guidelines. By employing a learning collaborative, this group was able to keep members interested and engaged. During the first year of the COVID pandemic the collaborative also served as a vehicle to share timely information.

Reflection on modern methods: causal inference considerations for heterogeneous disease etiology

Molecular pathological epidemiology research provides information about pathogenic mechanisms. A common study goal is to evaluate whether the effects of risk factors on disease incidence vary between different disease subtypes. A popular approach to carrying out this type of research is to implement a multinomial regression in which each of the non-zero values corresponds to a bona fide disease subtype. Then, heterogeneity in the exposure effects across subtypes is examined by comparing the coefficients of the exposure between the different subtypes. In this paper, we explain why this common method potentially cannot recover causal effects, even when all confounders are measured, due to a particular type of selection bias. This bias can be explained by recognizing that the multinomial regression is equivalent to a series of logistic regressions; each compares cases of a certain subtype to the controls. We further explain how this bias arises using directed acyclic graphs and we demonstrate the potential magnitude of the bias by analysis of a hypothetical data set and by a simulation study.

Risk Factors for the Severity of Complications in Minimally Invasive Total Gastrectomy for Gastric Cancer: a Retrospective Cohort Study

Purpose

Minimally invasive gastrectomy is a promising surgical method with well-known benefits, including reduced postoperative complications. However, for total gastrectomy of gastric cancers, this approach does not significantly reduce the risk of complications. Therefore, we aimed to evaluate the incidence and risk factors for the severity of complications associated with minimally invasive total gastrectomy for gastric cancer.

Materials and Methods

The study included 392 consecutive patients with gastric cancer who underwent either laparoscopic or robotic total gastrectomy between 2011 and 2019. Clinicopathological and operative characteristics were assessed to determine the features related to postoperative complications after minimally invasive total gastrectomy. Binomial and multinomial logistic regression models were used to identify the risk factors for overall complications and mild and severe complications, respectively.

Results

Of 103 (26.3%) patients experiencing complications, 66 (16.8%) and 37 (9.4%) developed mild and severe complications, respectively. On multivariate multinomial regression analysis, independent predictors of severe complications included obesity (OR, 2.56; 95% CI, 1.02−6.43; P=0.046), advanced stage (OR, 2.90; 95% CI, 1.13−7.43; P=0.026), and more intraoperative bleeding (OR, 1.04; 95% CI, 1.02−1.06; P=0.001). Operation time was the only independent risk factor for mild complications (OR, 1.06; 95% CI, 1.001−1.13; P=0.047).

Conclusions

The risk factors for mild and severe complications were associated with surgery, indicating surgical difficulty. Surgeons should be aware of these potential risks that are related to the severity of complications so as to reduce surgery-related complications after minimally invasive total gastrectomy for gastric cancer.

Effects of Lung Function-Related Genes and TSLP on COPD Phenotypes

A weighted genetic risk score (GRS) based on 16 SNPs implicated in reduced lung function in both Japanese and non-Japanese populations was previously associated with the onset of COPD and asthma. We here examine the genetic impact of this lung function GRS on specific COPD phenotypes. A cohort of Japanese COPD patients (N = 270) underwent lung function testing followed by genotyping with allele-specific arrays for 16 SNPs as well as expression quantitative trait loci at TSLP (rs2289276, rs3806933). Lung function GRS scoring and two-step cluster analyses grouped patients into different COPD phenotypes based on gender, age, smoking index, %FEV₁ and lung function GRS. The genetic effect of TSLP on COPD phenotypes was also examined for interactions with the lung function GRS. A total of 270 participants were grouped into 5 clusters. The cluster with the highest levels of lung function GRS was characterized by moderate to severe airflow obstruction and the highest blood eosinophil counts. Regarding TSLP, an increased number of T alleles at both SNPs was found in the cluster characterized by moderate to severe airflow obstruction and heavy smoking (rs2289276, p value = 0.035; rs3806933, p value = 0.047) independent of the lung function GRS. A genetic susceptibility to impaired lung function carries an increased risk of developing COPD characterized by increased eosinophil counts and severe airflow obstruction while individuals with increased TSLP responses to external stimuli have an independent risk of developing severe airflow obstruction in the presence of heavy smoking.

Socioeconomic inequalities in low back pain among older people: the JAGES cross-sectional study

BACKGROUND

Low back pain is an important public health issue across the world. However, it is unclear whether socioeconomic status (SES) is associated with low back pain. This study determines an association between SES and low back pain among older people.

METHODS

We used cross-sectional data derived from the year 2013 across 30 Japanese municipalities. The survey was conducted between October 2013 to December 2013. Functionally independent community-dwelling older adults aged 65 and above (n = 26,037) were eligible for the study. Multilevel Poisson regression analysis with a robust variance estimator was used to examine the association between SES and low back pain. Self-reported low back pain in the past year was used as a dependent variable. Educational attainment, past occupation, equivalized household income, wealth, and subjective economic situation represented SES and were separately analyzed as independent variables, adjusted for covariates including age and sex.

RESULTS

The prevalence of low back pain was 63.4%. Overall, lower SES were more likely to suffer from low back pain compared with that for the highest. First, as for the educational attainment, the prevalence ratio (PR) (95% credible interval (CI)) for the lowest level was 1.07 (1.02-1.12). Second, as for the past occupation, the PR (95% CI) for the blue-collared workers compared with professionals was 1.06 (1.01-1.11). Third, as for the equalized household income, the PRs (95% CI) for lower middle and the lowest income levels were 1.08 (1.02-1.13) and 1.16 (1.10-1.23), respectively. Fourth, as for the wealth, the PRs (95% CI) for lower middle and the lowest wealth levels were 1.11 (1.04-1.19) and 1.18 (1.11-1.27), respectively. Fifth, as for the subjective economic situation, the PRs (95% CI) for lower middle and the lowest financial conditions were 1.18 (1.10-1.26) and 1.32 (1.22-1.44), respectively.

CONCLUSIONS

Significant socioeconomic inequalities were observed in low back pain among older individuals in Japan. Policymakers and clinicians must understand the nature of these inequalities.

Inverse-Probability-Weighted Estimation for Monotone and Nonmonotone Missing Data

Missing data is a common occurrence in epidemiologic research. In this paper, 3 data sets with induced missing values from the Collaborative Perinatal Project, a multisite US study conducted from 1959 to 1974, are provided as examples of prototypical epidemiologic studies with missing data. Our goal was to estimate the association of maternal smoking behavior with spontaneous abortion while adjusting for numerous confounders. At the same time, we did not necessarily wish to evaluate the joint distribution among potentially unobserved covariates, which is seldom the subject of substantive scientific interest. The inverse probability weighting (IPW) approach preserves the semiparametric structure of the underlying model of substantive interest and clearly separates the model of substantive interest from the model used to account for the missing data. However, IPW often will not result in valid inference if the missing-data pattern is nonmonotone, even if the data are missing at random. We describe a recently proposed approach to modeling nonmonotone missing-data mechanisms under missingness at random to use in constructing the weights in IPW complete-case estimation, and we illustrate the approach using 3 data sets described in a companion article (Am J Epidemiol. 2018;187(3):568-575).