Internet Use by Primary Care Patients - STFM

ORIGINAL ARTICLES

Internet Use by Primary Care Patients: Where Is the Digital Divide?

Robin L. Kruse, PhD; Richelle J. Koopman, MD, MS; Bonnie J. Wakefield, PhD, RN; Douglas S. Wakefield, PhD; Lynn E. Keplinger, MD; Shannon M. Canfield, MPH; David R. Mehr, MD, MS BACKGROUND AND OBJECTIVES: Internet-based technologies such as personal health records and patient portals are increasingly viewed as essential for enhancing patient-provider communication and patient-centered care. We examined how primary care patients use the Internet, particularly patient characteristics associated with Internet use. METHODS: We surveyed patients in five primary care clinic waiting rooms. Patients who had used email or the Internet in the past month (Internet users) were asked how often they used a computer for a variety of tasks. Participants who reported not using the Internet were asked about several potential barriers to Internet use. RESULTS: We approached 713 patients, and 638 (89.6%) completed questionnaires; 499 (78%) were Internet users and 139 (22%) were non-users. Lack of computer access and not knowing how to use email or the Internet were the most common barriers to Internet use. Younger age, higher education and income, better health, and absence of a chronic illness were associated with Internet use. After controlling for age and other variables, chronic illness was no longer associated with Internet use. CONCLUSIONS: Internet use was high among our primary care patients. The major factor associated with Internet use among patients with chronic conditions was their age. If older adults with chronic illness are to reap the benefits of health information technology, their Internet access will need to be improved. Institutions that are planning to offer consumer health information technology should be aware of groups with lower Internet access. (Fam Med 2012;44(5):342-7.)

H

ealth information technology (IT) for consumers, such as personal health records, secure electronic messaging, and transmission of medical data from home-based devices, can enhance both patient-provider communication and patient-centered care.1 Whether in the context of the patient-centered medical home (PCMH),2 accountable 342

MAY 2012 • VOL. 44, NO. 5

care organizations,3 or meaningful use according to the 2010 Accountable Care Act,4 Internet-based health IT for patients is increasingly considered essential. Consumer health IT can encompass a variety of functions, including secure electronic mail (email) communication between patients and providers, online appointment

scheduling, prescription refill requests, importing clinical data collected outside the provider’s office (eg, home blood pressure monitoring), providing web links to a variety of patient education materials, viewing all or portions of a patient’s medical record, and allowing patients to add information to their record. Moreover, Internet-based applications can provide opportunities for novel approaches to targeting chronic health problems. However, many have suggested that there is a digital divide limiting many patients’ use of health IT, in particular poorer, sicker, and older adults.5-7 When primary care practices embrace the principles of the PCMH, planning for online health IT for consumers includes many critical decisions, such as determining who may participate, what services to offer, and with whom patients may directly communicate.8 As the number and complexity of the functions that are provided increase, the cost of implementing and maintaining the system increases as well. Because much health IT is Internet-based, From the Department of Family and Community Medicine (Drs Kruse, Koopman, and Mehr and Ms Canfield), Center for Health Care Quality and Department of Health Management and Informatics (Dr D Wakefield), Department of Medicine (Dr Keplinger), University of Missouri, Columbia; and Center for Comprehensive Access and Delivery Research Evaluation (CADRE), Department of Veterans Affairs Medical Center, Iowa City, IA, and the Sinclair School of Nursing, University of Missouri, Columbia (Dr B Wakefield).

FAMILY MEDICINE

ORIGINAL ARTICLES

it is essential to understand primary care patients’ current use of the Internet. National measures of Internet use, or even use among those with disability or chronic disease in the general population, may not reflect the usage patterns of patients attending a primary care clinic.9 In previous research, patients in three urban primary care clinics were interviewed about their Internet use in 2003; 53% of this population had used the Internet or email in the past year, and only 33% used the Internet to search for health information.10 Attending college was associated with greater online searching for health information, while respondents of Black race were less likely to use the Internet to search for health information.10 To better understand the potential audience for one academic medical center’s implementation of a patient web portal, we sought to examine primary care patients’ Internet use, particularly patient characteristics associated with Internet use. We focused on primary care patients because they often have a long-term relationship with their primary care provider and thus may be more interested in features such as secure electronic communication. We present results from a patient survey we conducted in five primary care clinic waiting rooms to better understand how our patients currently use the Internet.

Methods

We conducted a cross-sectional survey of outpatients in the waiting room at four family medicine clinics and one general internal medicine clinic at the University of Missouri, Columbia. Medicaid patients comprise approximately 15% of the patients in these clinics. The Health Sciences Institutional Review Board at the University of Missouri approved the study. Clinic sessions were randomly selected after stratifying by clinic and time of day (morning, afternoon, or evening) to obtain a representative sample. Sessions were weighted by provider FAMILY MEDICINE

hours within each clinic and specialty. Because resident physicians see fewer patients than attending physicians in our clinics, their sessions were assigned half as many hours (2) as attending physicians’ (4). Other than this difference, we assumed a constant rate of patients per provider session. Based on the weighted sum of hours per session, each was assigned a target number of completed questionnaires so that clinic sessions with more providers had a higher questionnaire target. Surveys were conducted in February and March of 2008 during 60 clinic sessions in the morning (n=23), afternoon (n=32), and evening (n=5). Clinic sessions were assigned to two research assistants who approached patients waiting to see their providers. All patients were approached unless the research assistant was busy talking with another patient, or it was obvious that the patient was not at least 18 years old. Rarely, patients were excluded because they could not understand English, hear the research assistant, or read the questionnaire. We recorded these patients as declining the survey. After confirming that the potential participant was a patient in the clinic and was at least 18 years old, they were asked if they were interested in participating. Consent was verbal, and the survey was anonymous. Research assistants stopped collecting data for a session once the target was achieved. After the target number of questionnaires for a clinic was collected, the clinic was removed from the sampling frame. We classified participants as Internet users or non-users based on screening questions that asked about their email and Internet use during the previous month. Those who reported using either Internet or email were considered Internet users, and those who reported using neither were considered non-users. While it is understood that email is part of the Internet, we segregated these concepts in our screening questions, as people sometimes view email use as being separate from the Internet.

This is a common and standard practice.9 We administered different paper survey instruments to users and non-users. We asked Internet users how often they used a computer for a variety of tasks (eg, email, read about general health information, order prescription medications) and some basic demographic questions. In addition to the same set of demographic questions, we asked non-users about several potential barriers to Internet use (eg, no computer access, too expensive, vision difficulties). Questionnaire data were double entered and verified for accuracy. We used SAS for Windows v9.1 to read data files and perform all analyses. We compared demographic characteristics between the two groups using chi-square analysis. For categories that were ordinal, such as age groups, we used the MantelHaenszel chi-square. To determine which demographic characteristics were independently associated with Internet use, we entered variables into a logistic regression model with Internet use as the dependent variable. We re-coded some variables to reduce the number of categories and to simplify interpretation.

Results

We approached 713 patients, with 638 completing questionnaires (89.6%). Only 39 (5.5%) patients declined participation; an additional 36 (5.0%) were seen for their visit before the questionnaire could be completed. Of those who completed questionnaires, 499 (78%) had used either email or the Internet in the previous month, and 139 (22%) reported neither email nor Internet use. Of users, 97% reported using both email and the Internet. Our participants ranged from 18 to 92 years of age; 70 (11%) were age 70 or older. Internet users and non-users differed on several demographic characteristics (Table 1). Younger patients were more likely to be Internet users than older patients, and Internet use declined with age (Mantel-Haenszel chi-square P