Table 4.

Covariates for monitoring factors contributing to changes on the outcomes.

Setting	Measure	Covariate(s)	Explanations
Ambulatory	Blood pressure control	Change in hypertension pharmacotherapy Acute illness	Uncontrolled patients with no pharmacotherapy changes may be false positives Acute illnesses may cause a temporary hypertension, but patient is still considered in control
	Diabetes bundle	Individual bundle items Type of health insurance	Evaluation of individual bundle items may facilitate identification of outcomes to improve Type of health insurance may be associated with chronic disease management
	Laboratory test orders	CDS alerts accepted Lab tests covered per type of visit Patient visits	Alerts of appropriate lab test may be associated with lab orders Changes in health insurance coverage may affect volume of lab orders Patient visits may be associated with lab orders
	Time documenting in EHR after hours	Risk adjustment factor Patient visits	Risk adjustment factor may be associated with electronic documentation Previous visits may be documented during work hours
	Patient visits	Time documenting previous visits Type of health insurance	Increased documentation may decrease patient visits Type of health insurance may decrease patient visits
	New patient visits	Proportion of patients per top insurance providers	Loss of patients from top insurance may decrease the number of new patients
Hospital	ED visits	Not identified during interviews	Not identified during interviews
	ED LOS ED wait time	ED visits Provider-patient ratio Go live support personnel*	More ED visits may increase LOS and wait time Provider-patient ratio may be associated with LOS and wait time More personnel for go live support may increase efficiency by shortening the learning curve
	MRSA infections CDiff infections	Patients in isolation	More patients in isolation may decrease infection rate
	Abdominal hysterectomy infections Colon surgery infections	Number of suspected infection cases according to the CDC’s NHSN	Number of potential infections captured by the EHR may help increase identification of true cases
	Employee turnover	Employee age	Employee age may be associated with resistance to learning a new EHR potentially increasing employee turnover
	Readmission rate	Appropriate use of medication for heart failure	Adherence to care pathways for heart failure may be associated with decreased readmission rate

Source: Covariates with data available in electronic format identified by the authors in the qualitative analysis. Abbreviations: CDS: clinical decision support; EHR: electronic health records; ED: emergency department; LOS: length of stay; CDiff: Clostridium Difficile; MRSA: Methicillin-resistant Staphylococcus aureus; CDC: Centers for Disease Control and Prevention; NHSN: National Healthcare Safety Network. *Potential moderator.