Table 2.

Methodological Evaluation of Studies

StudyFundingPopulationIndependent VariableDependent VariableOther Methodological Issues
Gadd (2000)None specifiedSampling method: non random, sequential Sample size: not justified Response rate: 82%EHR was in the exam room for more than 6 months prior to data collection Unsure how many physicians used it or how they didSurvey for patient satisfaction attitudes, developed and validated Blinding of data collector: N/A Blinding of analyst: not specifiedPossible variable experience with EHR usage inside the room Potential confounders considered in analysis: patient age, gender, self-rated computer sophistication or computer use
Gadd (2001)None specifiedSampling method: nonrandom Sample size: not justified Response rate: 77% in pre- and in post-Time (preimplementation and postimplementation of EHR) Unsure how many physicians used EHR or how they didPhysicians’ perception; instrument validated Blinding of data collector: N/A Blinding of analyst: not specifiedNo confounders were considered in the analysis Postintervention data collection over 18 months period
Garrison (2002)HRSA grantSampling method: nonrandom for post-, random for pre- Sample size justified Clear eligibility criteria Comparison to a historical sample Response rate 64% in post-Time (pre/post) Describes if physicians used EHR inside roomQuestionnaire designed for this study, piloted, but no mention of validation Blinding of data collector or data analyst: not mentionedAddressed respondent bias Analyzed if associations with patients’ perception of physicians’ computer skills and patients’ personal use of computer No adjustments for confounders mentioned
Hsu (2005)Garfield memorial fundSampling method: nonrandom; PCP volunteers and their patients; convenience Sample size: not justified Eligibility criteria defined Patients’ response rate: 80%Physicians familiar with the system prior to the intervention Training happened between data collection times in post (100% physicians attended)Self-administered questionnaire, pre-tested; based on “Medical Outcomes Study” No mention of blinding in data collection or analysisConfounders considered in analysis: age, gender, self-reported health status, race/ethnicity, annual household income, education, previously seen the PCP
Johnson (2005)Multiple*Sampling method: nonrandom in post- Eligibility defined Control: random selection, historical Sample size justified Response rate 92% (post)Use of EHR assessed after the encounter by physiciansQuestionnaire based on a tool developed by Wissow el al to assess satisfaction No mention of blinding of data collector or analystConfounders considered in analysis: Physician sex and year of training, possible within physician correlations
Joos (2006)No fundingSampling method: nonrandom, convenient Sample size not justified Response rate 66% (physicians)EHR already in place before this survey done No direct assessment of use of EHR, but general questions (not after each encounter)Survey built after qualitative section, tested for face validity No blinding for data collection or analysisConfounders considered in the analysis: gender, computer skill, message basket use and years of practice (but not for the outcome of interest)
Rouf (2006)HRSAEligibility specified Sampling method: nonrandom Sample size not justified Response rate 61% of attending and 38% of residentsEHR already in placeQuestionnaires from Medical Outcomes Study–reviewed for face validity–and piloted No blindingConfounders considered in the analysis: done but not specified which ones
  • * Funding from Robert Wood Johnson Foundation, Agency for Healthcare Research and Quality, Maternal and Child Health Bureau, Health Resources and Services Administration-Department of Health and Human Services, National Center for Injury Prevention and Control.

  • EHR, electronic health record; HRSA, Health Resources and Services Administration; PCP, primary care physician.