Understanding Adoption of a Personal Health Record in Rural Health Care Clinics: Revealing Barriers and Facilitators of Adoption including Attributions about Potential Patient Portal Users and Self-reported Characteristics of Early Adopting Users

Abstract

Personal health records (PHRs) are important for improving patient care. An important prerequisite to realize benefits of PHR use is patient recruitment. To understand clinic barriers to adoption, we used Rogers’ Diffusion of Innovations theory to frame an examination of clinic staff perceptions of a new PHR and perceptions of likely patient portal users. Clinic staff reported many relative advantages and observable benefits of the PHR but also some distinct problems. Attributions about potential patient users included demographic, computer use, and personality characteristics staff expected in likely users. Analysis of patient survey data of early adopters compared to non-users revealed discrepancies between clinic staff expectations and early adopters’ self-reports. Implications for improving adoption of PHRs include ensuring compatibility with existing systems and avoiding recruitment biases.

Introduction:

Personal health records (PHRs) are important resources for improving patient care, fostering patient engagement in healthcare, and informing clinical decision making [1]. To optimize these benefits, an important prerequisite is that patients are informed about the PHR and recruited to use it. Many patient-related barriers have been identified to initial use including poor patient knowledge of PHR functionality [2]. A patients’ initial recruitment and enrollment to participate in a PHR is often by clinic staff at a clinic which they routinely attend. This characteristic of implementation makes clinic level barriers to PHR adoption important to understand to promote meaningful use. In this study, we examined clinic staff perceptions of a newly introduced PHR with a patient portal in two rural health clinics. In addition, we identified clinic staff’s perceived characteristics of patients whom they thought would be likely (or unlikely) users of the patient portal. We wanted to understand why some patients were recruited whereas others might never have been approached for recruitment. Within both clinics, a relatively small proportion of patients comprised a cohort of “early adopters” who were interested in and actively using the patient portal component of the PHR. To test the accuracy of the attributions we used patient survey data to examine whether early adopting users who completed the survey and non-users responded to questions in ways consistent with clinic staff characterizations of likely users or non-users. Understanding clinic staff perceptions of the PHR, attributions about likely users, and the accuracy of these attributions reveals factors associated with clinic level adoption of a PHR and efforts to recruit patients for the patient portal component. These results inform future implementation efforts for PHRs, particularly in rural health clinics.

Background:

The PHR used in this study was designed as a component of a larger study.[3] The PHR emphasized integration of information and included a clinic-facing electronic personal health record and a patient portal. The larger research team worked with a health information technology company (CaduRX) to design the system. The PHR was demonstrated and adopted by two primary care clinics in rural areas. At the two clinics at which the Health Record had been adopted, there was a slower than anticipated patient recruitment rate. Clinic barriers to patient enrollment in a patient portal include physicians who are poorly informed about PHR functionality.[4] Clinic staff may also have negative experiences with a PHR or have drawn conclusions about patients that make them unlikely to approach patients for recruitment. To enhance use of a PHR it is important to understand clinic level barriers that may make patients less likely to be recruited for use.

The current study had the following aims:

1. Characterize clinic staff perceptions of a newly introduced PHR.

2. Categorize staff perceptions with criteria theoretically linked to likelihood of adoption.

3. Understand staff perceptions about the “kind of patient” they predict to be likely to use the PHR patient portal to reveal barriers to recruitment.

4. Test whether staff predictions about likely patient portal users are similar to the self-reported characteristics of early adopting users of the patient portal.

Understanding incorporation of a new system into an existing structure is a challenging yet predictable process. In this study, we were focused on the adoption phase- understanding how the adoption of the PHR might be helped or hindered within the specified clinics. Rogers’ Theory of the Diffusion of Innovations provides a framework for understanding what innovations will be adopted within a setting.[7] Rogers predicts adopted innovations are (1) relatively advantageous over existing systems, (2) compatible with existing systems, (3) have observable benefits (4) trialable (able to be modified), and (5) low in complexity, or, easy to understand and use. Innovations that fit such a profile are likely to be adopted and incorporated into the existing system.[7] In the current study, the diffusion of innovations frame contributes to understanding early use patterns of the PHR in the test clinics.

Rogers also identifies the importance of agents of change – individuals who drive change by influencing those around them to uptake the innovation. Change agents have personal characteristics and communication skills that contribute to their ability to exert social influence. For a new technology, such as the PHR implemented by these clinics, change agents might value technology, recognize it as useful, and encourage colleagues and patients to identify useful functions through direct exhortations or indirectly by modeling use. Change agents are likely to exert influence on those who they believe to be likely adopters and potentially change agents may influence others with their views about likely adopters. Within a clinic, change agents might promote recruitment or suppress recruitment based on opinions about the PHR and attributions about potential users thus it is important to understand individual clinic staff members’ perceptions of the PHR and to understand which patients staff might approach for recruitment as the perceptions of some staff members may influence the likelihood of adoption more than others.

Understanding what characteristics clinic personnel ascribe to potential patient users provides a roadmap for understanding which patients the clinic personnel are likely to encourage to use the patient portal. Attributions about the “kind of person” who would engage in a particular behavior are influenced by lay theories of personality and judgments about the desirability of engaging in a behavior.[8, 9] The current study provides insights into the attributions of clinic personnel about patients willing to use the patient portal and about patients expected to be disinterested in or to decline to use the patient portal. This study uncovers reasons why clinic personnel might choose to recruit specific patients to use a PHR-providing valuable information for other groups interested in implementing a PHR for patient care or research purposes.

Methods:

Primary care clinics in two distinct rural communities of Utah were enrolled in the PHR during the study period. The rural communities served by these study clinics range in population size from about 6,500 to 22,500 individuals. All study procedures were approved by the Institutional Review Board at the University of Utah.

Interview Methods:

A semi-structured interview guide was developed to allow the interviewers to identify experiences with the PHR, features enjoyed, and experiences recruiting participants. The methods for interview construction were consistent with recommendations by Patton.[10] The interview was designed to understand features of adoption and reveal clinic staff members lay theories about potential users. The guide was tested on research staff and all interviews were conducted using uniform procedures with the interview flow guided by the participants. The interview questions related to features of the PHR and characteristics of potential patient users. The interviews were conducted by two research team members at the clinics during a time convenient for the interviewees. Clinic staff participants included physicians, providers, office managers, medical assistants, and receptionists. A total of 9 in-depth interviews were completed. Interviews were conducted two-on-one (two research team members and one clinic staff member) so that clinic personnel could feel free to express their ideas without the possible pressure of having other clinic staff present to avoid the potential impact of problems relating to underlying group dynamics or relationships. Interviews were recorded and transcribed. All identifiable information was left out of the recordings and the recordings were stored on a secure drive.

Survey Methods:

A telephone survey of patients from participating clinics was conducted to understand characteristics of the clinic patient populations potentially predictive of patient portal adoption. Specifically, patients were queried about experiences within the clinic (e.g., wait time and frequency of visits), patient satisfaction with communication with their doctor, patient technology use (email at home), and patient activation and health history. Eligible patients first received a letter from their clinic explaining the purpose of the survey. Letters were followed by a phone call that further explained the survey and recruited the patient to participate. Patients could refuse participation at any time. Patients who had been diagnosed with a chronic illness (specifically, diabetes, congestive heart failure (CHF), chronic obstructive pulmonary disease (COPD), high cholesterol, and hypertension) were recruited 3:1 for every patient who had not been diagnosed with a chronic illness. Patients with chronic illness were oversampled because these patients were believed by the study team to be more likely to benefit from using the PHR given their health management and clinic access needs. In addition, participants were asked several questions about their access to and use of technology in general and for managing their health. The survey was composed of pre-existing, validated survey tools (e.g. Patient Activation Measure; PAM[11]) as well as questions developed by the researchers to collect specific project related items of interest. Data were stored on a secure drive.

Interview Analysis Methods:

The interview transcriptions were initially iteratively reviewed by members of the research coding team who included a health psychologist, an informaticist, and a data scientist/epidemiologist. Using an analysis approach consistent with recommendations by Patton,[10] a coding guide was developed using Roger’s Theory of Innovations [7] for coding comments related to the PHR. Patient user attributes were coded consistent with patient user characteristics and clinic staff attributions. Direct quotes within the text that reflected topics of interest were marked a process Patton calls “pre-coding”.[10] Then, the research team coded one interview together to reach consensus about which specific quotes would be assigned to which categories. Other transcripts were coded by one coder and then reviewed by a second coder. Any disagreements were discussed by the full coding team. PHR related quotes were categorized as endorsing or negating dimensions of Roger’s Diffusion of Innovations Theory. For example, the quote, “”I like CaduRX…it’s really… user friendly.” was coded “low in complexity” whereas a quote about one function of the health record, “It just has a whole lot of steps, I’ve seen it done easier” was coded “high in complexity”. Patient related quotes were coded as referencing potential patient users (and non-users) computer use, health maintenance, attributions about personality, and patients’ need for convenience to reveal what staff expected in likely patient adopters of the patient portal.

Survey Analysis Methods:

The survey data were exported into the SPSS statistical analysis software package.[12] Early adopting users were identified by use patterns that tracked patient portal use and a variable was created categorizing survey respondents as “early adopters” or not. Early adopters used the patient portal at least one time on a different day from their initial recruitment log in during the first 9 months in which the patient portal was available within the clinics. Comparisons between the survey responses of early adopters and non-users were conducted using t-tests.

Results:

Clinic staff characterized the PHR in both positive and negative ways. The PHR was considered advantageous, thought sometimes incompatible with existing systems. Benefits were often observed, but some reporters indicated a lack of observable benefits. The PHR was reported low in complexity and easy to use. There were few mentions of trialability although some participants asked interviewers for improvements and referenced hopes for imminent changes suggesting that the system was generally viewed as modifiable. Clinic personnel interviews related positive experiences with the PHR within the clinics as well as a number of difficulties. Additional themes emerged in the analysis indicating factors that clinic staff found positive about the system (e.g., features identified as relatively advantageous) (see table 1 for summary) and factors they found to be relatively negative (e.g., problems with compatibility with existing systems) (see Table 2 for summary).

Table 1.

Interview Themes related to Likelihood of Adoption Based on Rogers’ Theory of Diffusion of Innovations

Adoption Theme Category	Interview Theme
Relatively Advantageous	Convenience of medication prescribing and clinic notes system
	Fraud prevention (no prescription pad, direct secure communication with pharmacy, patient information such as drug dependency easily accessible)
Compatible with Existing Systems	Seamless integration with pharmacy system (in one clinic)
	Integration with dictation for support claims
Observable Benefits	Necessary procedures for diagnosis or problem available to patient care team at start of appointment
	Fewer patient calls
	Improved patient tracking (frequency of visits, screening test intervals)
	Once information is entered, saves time at appointments
Triable (modifiable)	Expectation that enrollment glitches would be fixed
	Recording desired features to report to PHR development team
Low in Complexity	Intuitive for users
	Access needed information from any computer (e.g., notes, prescription information, patient reported allergy history)

Table 2.

Interview Themes Related to Barriers to Adoption Based on Rogers’ Theory of Diffusion of Innovations

Problem Type	Problem Description
Not Relatively Advantageous	Time required to log in, enter data and populate health maintenance tools is frustrating
	Speed of internet connection in clinic impacts functionality
	Office staff can allow prescriptions, errors happen when MD just authorizes without double checking information (e.g., dose)
Not Compatible with Existing Systems	Billing problems – (e.g., must enter insured as patient even if they are not)
	Patient labs are not integrated in system
	Glitch getting information (e.g., SSN) to appear in patient record
	Clinic Role impacts entry permissions – sometimes medical assistant cannot print or enter needed information
	Integration with pharmacy systems differs by pharmacy
Lack of Observable Benefits	Patient entered data in health history too lengthy and detailed for MD use
	Many features not used by clinic staff
Not trialable (not modifiable)	Unclear and slow process for reporting and resolving technical glitches- like repeated patient password problems on enrollment
	Changes occur in negative ways e.g., familiar settings are changed to unfamiliar
High in Complexitiy	System features difficult to understand

Relatively Advantageous:

Interview Results Relating Positively to Adoption

The clinic facing component of the PHR was regarded as having advantages over previous systems. Particularly valued components were clinic notes systems and prescription ordering features. One participant indicated, “it’s nice to be able to see, if we wrote a note, who wrote the note, scheduled the patient. I do like that part of it…”. Ordering electronic prescriptions was discussed extremely positively in 8 of 9 interviews, “They (patients) know (the prescription) is going right to me and it’s even going to go to me if I’m at home. So they really like that and it does cut down on the delays (in filling prescriptions)”, indicating that this advantage was perceived to help the clinic and individual patients. Prevention of fraudulent prescriptions was considered of particular value. One interviewee noted that the PHR eliminated the need for prescription pads and reported a related anecdote about a stolen pad resulting in multiple fraudulent prescriptions tied to the clinic, “(the) patient stole the pad he had us going to five different pharmacies every day of the week Monday, Tuesday, Wednesday, Thursday, and Friday to a different pharmacy in (nearby town)….and getting Loratab and getting 30 a week but five times.” Interviewees noted that filling prescriptions by telephone was particularly prone to fraud, making the electronic request feature a particular benefit, “…even Loratab can be called in over the phone. All those narcotics like that….I mean people know my name but unless they know my voice, most of the time you will get a voice activated answer machine that says this is (staff) from Dr. (’s) office refill John Does prescription of Loratab # 30, and that’s the end of it”. One staff member reported being “called on the carpet” by clinic supervisors for issuing a narcotics prescription on her day off, which was revealed to have been fraud. The direct, secure connection from clinic to the pharmacy was seen as protective for the clinic, for staff members themselves, and clearly advantageous to the previous system featuring a prescription pad and/or telephone access.

Many observable benefits of the PHR were related to workflow including improvements in tracking the frequency of patient visits. Clinic personnel reported that patient tracking prior to the introduction of the PHR was difficult, particularly given the rural nature of the community, “that was very helpful because sometimes doctors will just repeatedly oh we see so and so and especially in a small community everybody knows everybody so if you’re not paying attention we’ve just seen him, well no it’s been two and a half years since we’ve seen him”. Clinic personnel reported that visits in the clinic were more efficient and better coordinated across the patient care team when the PHR was used. As one physician reported, “there isn’t a patient that walks in that I don’t look at (the problem) list and of course I’ve created that from my own practice so I’ve created that list. But it pretty much tells me and my (staff) what to do next. My nurse will look at the list that’s printed from the last visit and if there’s something she can do before hand like an EKG or a chest x-ray, she knows it’s time on someone, it’s that yearly, she’ll just do it before I even see the patient. That’s really helped with the flow…. I don’t even have to think real hard (be)cause it does it for me.” The physician cited the PHR tools as contributing to workflow and to reducing cognitive burden of routine patient care, ostensibly freeing her to focus on more complex patient care issues.

The remaining categories theoretically related to greater likelihood of adoption, (compatibility with existing systems, trialability and low complexity) were mentioned less pervasively in interviews. The prescribing system was indicated by one participant to be seamlessly integrated with prior systems, “I didn’t notice it any different from if the patient requested it or if we just put it in from the fax from the pharmacy”. There were minimal mentions of the PHR as trialable. Multiple participants did indicate that they expected technical issues to be resolved as one participant mentioned, “things that can be fixed if they have the time” and another indicated writing down desired features to provide to the PHR development team. There were several indications that the system was perceived to be low in complexity with participants describing the PHR as easy to use, “user friendly” and as operating very smoothly, “It is very slick. It is very nice”. The PHR system was regarded as modestly compatible and trialable but quite easy to use.

Not Advantageous With a Lack of Observable Benefits:

Interview Results Relating to Clinic Level Barriers to Adoption

Incorporating the EHR was not viewed as a uniformly positive experience. Ways in which the system was described as not advantageous over existing systems included time costs particularly on high volume patient days, “you know when you’re seeing 25 patients in a day, you sit down and record those things that becomes a time issue. So it’s been done erratically.”. In one clinic, one physician was described as “hating to log in all the time” indicating that even fairly basic system use requirements were considered burdensome by some individuals at some times. One clinic had a particularly slow internet connection which blunted the potential sense of advantage. The system was also described as potentially contributing to errors, “the other problem is when the girls in the front put the information in the computer, because they’re not nurses and they’re not doctors, a lot of times they’ll mess up the information that they put in the computer and then the pharmacy will call us and say are you sure you want to, you know, double this or whatever”. There were no participants who described the system as entirely without advantage over previous systems but there were also no participants who did not report some challenges with the system.

Several problems of compatibility with existing systems were reported. Billing was a particular headache with the insured person (who might be a patient spouse or other relative) forced to be entered as the “patient” even when they were not. The PHR could not be used to retrieve lab results because those were stored in another system not yet integrated with the PHR. Medical assistants reported difficulties with permissions, including printing patient medication lists. One clinic had a problem relating to integration with a particular pharmacy with the pharmacy frequently indicating they had not received the medication order. This lack of compatibility emerged more strongly in data analysis than demonstrations of compatibility, suggesting that these may have been particularly important barriers to adoption.

A lack of observable benefits of the PHR was reported by some participants. These concerns included lack of usefulness of data entered by the patient into the PHR and viewable by clinic staff, “I don’t find it super useful. When they come they’ve got a lot more things than I ever wanted to know. And they have more things than I want on that list. Like I had the measles at age 4.5 you know those things I just erase because they just clutter up my list.” Conversely, another staff member found the patient-entered history to be useful and encouraged patients to enter the information in detail, “If you had tonsillitis, you had your tonsils out, you know, we don’t know you don’t have your tonsils out. Put the date down. They go, oh, I didn’t think about that” suggesting that at times clinic staff may have been at cross-purposes in encouraging use. Another way in which lack of benefits was revealed was by participant statements that they had never used particular features including information entered by patients in the portal and health maintenance tools, “I haven’t used (health maintenance tools). I’m not sure why I haven’t.”

For the remaining categories, trialability and complexity, participants rarely reported that the system was difficult to use. A few indicated that they had trouble with very basic functions (entering passwords) suggesting a general lack of familiarity with computer-based systems. Some participants reported negative changes, “sometimes, when we’ve wanted (the PHR development team) to change something and they’ve said they would change something then it might change for a couple of days and then it suddenly changes back….like for example (we) wanted the prescriptions to say when they were denied and for what reason they were denied”. Thus, by some participants the system was viewed as changeable but as likely to be “changed back” as initially modified.

Clinic Staff Descriptions of Patient PHR Users

Clinic personnel were asked about the kinds of patients they thought were interested in using the patient portal. In response, participants reported expectations related to users and, conversely, characteristics of non-users. The descriptions of such patients were diverse, but consistent themes emerged related to computer use, health and health management, demographics, attributions about personality, and factors related to patient convenience.

Potential patient portal users were described as technical capable, “computer literate”, “computer savvy”, and enjoying computer use, e.g. as patients who “like to play on the computer”. Patient users were also characterized as meeting specific health or health management related criteria. Clinic personnel described users as chronically ill “They have problems that keep ‘em coming back like diabetes or cholesterol or something that keeps ‘em coming back on a regular basis”. One interviewee described patient users as “people who really, truly care about their health are the ones who are using it the most”. Potential portal users were often described as desiring the convenience of portal use, “people who work difficult hours” and “get controlled substances regularly” as these patients were seen as particularly benefitting from the access to the patient portal from any internet accessible computer.

The demographic characteristics of potential patient portal users were described as “younger” though not “very young”, and also as typically “over 50” with this age estimate described as due to the likelihood of patients over 50 having a chronic illness. Over 50 is relatively young for the patient population in the clinic. Attributions about users often centered on personality characteristics. Patient users were described as “the aggravating patients”, patients who frequently called the clinic, and having organized personalities, “really anal about liking to know every detail of their stuff. People who tend to be extremely organized at home are the same people that like this”. Thus, the general picture of users according to clinic personnel was relatively young, chronically ill, motivated to care for their health, and focused on healthcare resources.

Clinic Staff Descriptions of Patient non-Users of PHR

When describing patient users and potential users, attributions about non-users emerged. Non-users were characterized as computer illiterate or completely inexperienced “don’t have a computer, don’t understand those things”. Older non-users were contrasted with users on social support dimensions related to computer use, “there have been some people….they’re older who have kids that are, you know, computer literate and they help them do that. But the ones who never, who don’t have kids and stuff like that, they’re interested but they don’t have a computer, they have no clue how to, you know what I mean, I’ve never used a computer” Some clinic personnel believed non-users to be “non-compliant” with medical regimens. Non-users were frequently characterized as fearful, “afraid of computers” and “don’t trust their information is going to be safe”. Non-users were also described as younger, very young, and not chronically ill.

Patient Survey Results Characterizing Early Adopting Users: Patient survey data was used to examine characteristics of patient portal users. A group of 39 patients who completed the survey were identified as users of the patient portal. Because the survey was conducted early in the course of adoption, these 39 patient users might be characterized as early adopters of the PHR. Of the 39 early adopters, 22 were between age 45 and 64. An additional 14 were 65 or older and 3 were under 44. The gender breakdown was fairly even, with 21 female early adopters and 18 male. Most of the early adopters had a computer at home (35 of the 39). In addition, 32 had a personal email account and 35 had internet access at home. The mean frequency of email contact was 4.53 falling in between “several times a week” and “at least once daily” on a 5 point likert scale. In addition, comfort with the internet was high for these early adopters, with the mean comfort level of 3.32 indicating that the average early adopter was “comfortable” or “very comfortable” using the internet. The early adopters seemed to report relatively positive experiences with their healthcare clinic. Mean early adopter ratings of providers listening carefully, explaining information in a way that was easy to understand, and spending enough time with the patients were between 5 and 6 on a likert scale indicating that these early adopters thought that their providers behaved this way nearly an average of almost always to always. In addition, the early adopters were extremely satisfied with their providers with a mean satisfaction rating of 9.13 on a 10-point scale. A perfect ten represented “the best possible provider”.

Early adopting patient PHR users were compared to the 365 non-user survey respondents within the clinics in which the PHR was implemented who completed the survey. When possible, self-reports of early adopter patient users were compared to non-users in conceptual domains mentioned by clinic personnel as potentially related to “the kind of person” who would be interested in using the patient portal. Comparisons included age, access to computers, health management, and experience with the clinic. The early adopters were not different from non-users in age, gender, having a chronic illness, reports that the provider explained or listened carefully, spent enough time with them, or satisfaction with the provider. This group of early adopting users were significantly different from their non-user counterparts on a single dimension related to computer use, comfort with internet use (t (391) = −2.74 (p< .01)). On a 5-point scale item of comfort with internet use, with higher numbers indicating greater comfort, users early adopting users reported a mean of 3.32 whereas their non-users counterparts reported a mean of 2.81. Notably, early adopters were marginally more likely (p-values less than .10) to have access to a computer at home, have an email account for personal use, and to report access the internet regularly. Potentially, a larger sample of adopters would allow these dimensions to reach statistical significance. As is, the data suggest a trend toward greater access to, and regular experience with, computers and the internet.

Discussion:

The in-depth clinic interviews revealed that the clinic experience with the PHR was positive. The PHR was considered easy to use and providing important advantages and observable benefits. In particular, the clinic staff related vivid examples relating to the pitfalls in prescribing controlled substances that could be avoided by using the PHR. The salience of the relayed anecdotes suggest that this was a particular concern for the staff and a close-held benefit. A number of barriers may have truncated adoption of the PHR within the clinic. Though most concerns were minor and could be modified in the future (e.g., integration of labs), the failures of compatible and difficulty making changes may have impaired adoption rates. Research and care teams promoting clinic level adoption might consider implementing systems highly compatible with existing systems. When this is not possible, teams should demonstrate trialability by creating specific, transparent procedures for communicating modification requests and tracking modifications made. In addition, some participants may have been agents for change in the positive – but others were fairly negative about the system. Teams should use the tendency for groups to self-organize around agents to change to encourage PHR champions within the clinic.

Potential users of the patient portal were viewed extremely positively – they were seen as technically capable, invested in their own health maintenance (e.g. “people who really, truly care about their health”) and organized. In contrast, non-users were seen primarily as fearful, old, and having less social support. Potentially, clinic personnel viewed potential patient portal users positively because they viewed the PHR positively and thus proscribed positive attributes to PHR users. Even pejorative terms attributed to PHR users (e.g. “anal” or “aggravating”) were couched in positive explanations of the patient users as active managers of their health. Negative views about non-users may have been influenced by bias against older people.[13, 14] In contrast to the generally positive descriptions attributed to portal users older users were characterized somewhat negatively, one interviewee stated that older users did not use all the aspects of the PHR, such as requesting refills using the system, “It’s hard to break old habits and (they) have been used to calling on the telephone when they want something and to change that, that’s the old dog new trick…” Thus, users were considered active, informed patients unless they were old.

Examining self-reports of patient early adopting users compared to non-users indicated that clinic personnel opinions bore little validity. Early adopting users were more comfortable with internet use but other computer related access points were only marginally related. Consistent with the confused description of the age of potential portal users (e.g., older, younger, not very young) the patient users were not different in age from the non-users. No differences were found across health management domains. There was little indication that early adopting users were more aggravating, though this was not directly measured but by proxy, early adopters were as satisfied with their care as non-users and equally satisfied with time spent with providers.

Clinic personnel were invested in drawing in additional patient users. This enthusiasm of these agents of change may have occasionally been detrimental for other clinic personnel as in the case the medical assistant encouraging entry of a detailed patient history, the same type of history reported as of very limited use by a physician user. During early adoption a flexible dynamic is expected. Eventually, new norms will be established by the user group. Based on Rogers’ work, it would be anticipated that the PHR would continue to expand across clinic patients but the hitches in compatibility must be addressed.

It is important to note that the patient survey data was not designed to answer personality related questions about “the kind of patient” who was interested in using the patient portal, but many of the domains were adequately represented. Much of the information reported did address concepts cited by the interviewees. Implications of the current work suggest that when using clinic staff to promote use of a PHR it is important to consider attributions, biases, and opinions of these agents. For example, agents of change with a strong age bias may have been less likely to invite older patients to enroll in the PHR due to the perception that such patients would be uninterested. Promotion strategies should focus on inclusiveness and overcoming biased opinions to allow broad groups of patients to benefit from PHR use. Given the survey evidence, many of the biases may be unwarranted. Further research is needed to assess attributions formed that may influence PHR patient recruitment rates. Finally, results suggest that prevention of fraud and workflow improvements are highly valued benefits in rural health clinics and can be influenced by incorporating a PHR.

Table 3.

Descriptions of Patient Users by Health Care Providers, Clinic Staff

Patient User Categories	Description
Computer Use	Computer literate/computer savvy
	Frequent computer users
Health/Health Management	Chronic health problems (e.g., cholesterol, diabetes)
	Patients who regularly take controlled substances
	Patients who are adherent to medical regimens
Patient convenience	Patients who like to handle things at home
	Those who work difficult hours
	Frequent travelers
	Patients who don’t like to call the clinic office
	People who value free time
Attributions about user/potential user personality/Demographics	Patients who care about their health
	Older people with kids to help them use the computer
	Patients 50 and above
	Highly organized patients
	Patients demanding of clinic resources
	Proactive patients

Table 3.

Descriptions of Patient Refusers by Health Care Providers, Clinic Staff

Patient User Categories	Description
Computer Use	Patients without computers
	Patients not computer literate
Health/Health Management	Very young patients
	Patients free of chronic illness
	Non-compliant patients
Attributions about non-user personality/Demographics	Patients who don’t trust computer
	Patients who don’t trust information will be safe
	Patients afraid of computer