Clinical data mining: challenges, opportunities, and recommendations for translational applications

Table 2 The main heterogeneities in clinical practice for data mining

	Source of heterogeneity		Attributes
Participants	Demographic characteristics		Spatial heterogeneity; Time heterogeneity; Space–time heterogeneity
	Phenotype
	Genotype
	Behavioral characteristics and social factors
Interventions	Proficiency in professional skills		Diversity of therapeutic regimen (monotonically improvement); Diversity of clinical practice guidelines
	Nursing quality
	Medical quality
	Accessibility of medical devices
Outcomes	Type of the outcome: primary and secondary outcomes, side effects, disease progression		Subjectivity: doctor subjective report and patient self-report
	Definition of the outcome: including binary and continue with cut-off		Objectivity: diagnostic report (imaging, pathology, laboratory tests)
	Observation duration		Time effect: timeliness or lateness of outcome occurrence time
Comparisons	Case–control	PS Matching	Post-hot randomization
	Case–control	PS Weighting (IPTW, SMRW)	Non-randomization
	Cohort studies	Instrumental variable	Randomization-like
	Randomization		Randomized controlled trial

IPTW inverse probability of treatment weighting, SMRW standardized morbidity ratio weighting

ISSN: 1479-5876