Increasing Response Rates When Surveying Medical Practices: An EHR Survey Example

To determine electronic health record (EHR) use and adoption rates, Ohio State University conducted a mixed-mode sample of 4,843 Ohio primary care physicians, dentists, pediatricians, medical specialists, and nurse practitioners/nurse midwives. The sample frame consisted of Medicaid's 2010 provider list and the National Provider Identifier Registry for Ohio. A staged process of postcard notification, e-mail, mailing of the instrument for fax response, follow-up postcard and e-mails, and an option of instrument delivery was implemented. The survey aimed to determine EHR uptake, barriers to adoption, and likelihood of EHR adoption. Fielding was scheduled for 6 weeks. After 3-weeks, 6.4% of contacts returned a completed survey. Given low response rate (RR), we adjusted the protocol and initiated a telephone campaign to urge survey completion. Options given included: phone interviews, return by fax, return by e-mail, and return by postage – inclusive faxing of non-respondents included a cover letter and instrument. This was followed by phone contact and resulted in a 19.3% RR. Given a mixed-mode collection, a judgement poststratification estimator of having an EHR was employed. The mixed-mode was most effective for specialists, primary care physicians, and dentists, and least effective for nurse practitioners/nurse midwives. We conclude that medical professionals are more likely to respond to mixed-mode surveys. The survey found that small practices report being financially inhibited from fully implementing EHRs and that assistance will be needed to meet national recommendations for EHRs – especially in small rural practices.