Adaptive Modeling: An Approach for Incorporating Nonlinearity in Regression Analyses

Three machine learning algorithms and their utility in exploring risk factors associated with primary cesarean section in low-risk women: A methods paper

Machine learning, a branch of artificial intelligence, is increasingly used in health research, including nursing and maternal outcomes research. Machine learning algorithms are complex and involve statistics and terminology that are not common in health research. The purpose of this methods paper is to describe three machine learning algorithms in detail and provide an example of their use in maternal outcomes research. The three algorithms, classification and regression trees, least absolute shrinkage and selection operator, and random forest, may be used to understand risk groups, select variables for a model, and rank variables' contribution to an outcome, respectively. While machine learning has plenty to contribute to health research, it also has some drawbacks, and these are discussed as well. To provide an example of the different algorithms' function, they were used on a completed cross-sectional study examining the association of oxytocin total dose exposure with primary cesarean section. The results of the algorithms are compared to what was done or found using more traditional methods.

Mothers' and fathers' views of family management and health-related quality of life for young adult survivors of childhood brain tumors

PURPOSE

To examine associations between fathers' and mothers' appraisals of family management and physical and emotional health-related quality of life (QOL) for young adult survivors of childhood brain tumors.

DESIGN

Cross-sectional.

SAMPLE

47 mothers and 39 fathers (39-67 years old); 47 survivors (18-33 years old).

METHODS

Analyses evaluated relationships among family management (Survivor's Daily Life, Condition Management Ability, Condition Management Effort, Family Life Difficulty, View of Condition Impact, Parental Mutuality), quality of life, and parental role.

FINDINGS

Except for Parental Mutuality, family management ratings were not significantly different for mothers and fathers, and parental views of survivors' physical and emotional QOL improved with better family management. Parental role moderated associations between physical and emotional QOL and Survivors' Daily Life and between emotional QOL and Condition Management Ability, Condition Management Effort, and View of Condition Impact.

IMPLICATIONS FOR PSYCHOSOCIAL PROVIDERS

Assess and address survivor QOL through family management from multiple perspectives.

Using baseline target moderation to guide decisions on adapting prevention programs

Tom Dishion, a pioneer in prevention science, was one of the first to recognize the importance of adapting interventions to the needs of individual families. Building towards this goal, we suggest that prevention trials be used to assess baseline target moderated mediation (BTMM), where preventive intervention effects are mediated through change in specific targets, and the resulting effect varies across baseline levels of the target. Four forms of BTMM found in recent trials are discussed including compensatory, rich-get-richer, crossover, and differential iatrogenic effects. A strategy for evaluating meaningful preventive effects is presented based on preventive thresholds for diagnostic conditions, midpoint targets and proximal risk or protective mechanisms. Methods are described for using the results from BTMM analyses of these thresholds to estimate indices of intervention risk reduction or increase as they vary over baseline target levels, and potential cut points are presented for identifying subgroups that would benefit from program adaptation because of weak or potentially iatrogenic program effects. Simulated data are used to illustrate curves for the four forms of BTMM effects and how implications for adaptation change when untreated control group outcomes also vary over baseline target levels.

Sex-specific Associations Between Serum Hemoglobin Levels and the Risk of Cause-specific Death in Korea Using the National Health Insurance Service-National Health Screening Cohort (NHIS HEALS)

Objectives

The purpose of this study was to determine the associations between blood hemoglobin (Hgb) levels and the risk of death by specific causes.

Methods

Using the National Health Insurance Services-National Health Screening Cohort (n=487 643), we classified serum Hgb levels into 6 sex-specific groups. Cox regression analysis was used to analyze the associations between Hgb levels and the risk of cause-specific death.

Results

Hgb levels in male population showed a U-shaped, J-shaped, or inverse J-shaped association with the risk of death from ischemic heart disease, acute myocardial infarction, liver cancer, cirrhosis and chronic obstructive pulmonary disease (COPD) (all non-linear p<0.05; hazard ratio [HR]; 95% confidence interval [CI]) for the lowest and the highest Hgb levels for the risk of each cause of death in male population: HR, 1.14; 95% CI, 0.98 to 1.34; HR, 2.87; 95% CI, 1.48 to 5.57; HR, 1.16; 95% CI, 0.96 to 1.40; HR, 3.05; 95% CI, 1.44 to 6.48; HR, 1.36; 95% CI, 1.18 to 1.56; HR, 2.11; 95% CI, 1.05 to 4.26; HR, 3.64; 95% CI, 2.49 to 5.33; HR, 5.97; 95% CI, 1.44 to 24.82; HR, 1.62; 95% CI, 1.14 to 2.30; HR, 3.84; 95% CI, 1.22 to 12.13, respectively), while in female population, high Hgb levels were associated with a lower risk of death from hypertension and a higher risk of death from COPD (overall p<0.05; HR, 1.86; 95% CI, 1.29 to 2.67 for the lowest Hgb levels for hypertension; overall p<0.01, HR, 6.60; 95% CI, 2.37 to 18.14 for the highest Hgb levels for COPD). For the risk of lung cancer death by Hgb levels, a linear negative association was found in male population (overall p<0.01; the lowest Hgb levels, HR, 1.17; 95% CI, 1.05 to 1.33) but an inverse J-shaped association was found in female population (non-linear p=0.01; HR, 1.25; 95% CI, 0.96 to 1.63; HR, 2.58; 95% CI, 1.21 to 5.50).

Conclusions

Both low and high Hgb levels were associated with an increased risk of death from various causes, and some diseases showed different patterns according to sex.