Now Reading: Impact of Health Literacy on access and use of a Personal Health Record

I’ll let the authors words describe the significance of this study:

this is the first study to assess real-world use of an industry-leading patient portal (Kaiser Permanente’s My Health Manager) in a large, ethnically, culturally, socioeconomically, and educationally diverse sample of diabetes patients with a range of self-reported health literacy skills.

It is also a great companion to the excellent post and review about the use of Personal Health Records in Washington, DC, by John Moore. I have written a lot on this blog ( a lot ) about the potential of personal health records for vulnerable populations, based on observations and past experience. We are now entering the era of actual data and use, which is going to make our work even stronger.

In a nutshell, Kaiser Permanente patients with diabetes were surveyed regarding the potential for disparities in their condition management and outcomes, and crossmatched with their use of Kaiser Permanente’s My Health Manager as of January-December, 2006. Important to keep this in mind as a challenge of health services research – 2006 is a long time ago in internet years. At the same time, there hasn’t been a population of this size attached to a functional personal health record anywhere in health care, which is why this information is so valuable.

It is worth quoting here how the researchers measured health literacy, it will help other designers of personal health records (and health systems, really):

This instrument asks about problems due to reading, understanding, and fill- ing out forms, not due to poor vision: (a) ‘‘How often do you have problems learning about your medical condition because of difficulty understanding written information?’’; (b) ‘‘How confident are you filling out medical forms by yourself?’’; (c) ‘‘How often do you have someone like a family member, friend, hospital or clinic worker or caregiver, help you read Kaiser health plan materials (such as written information about your health or care you are offered)?’’

The results and their presentation are on the complicated side, in my opinion, however, I did my best to summarize using these two slides.



In 2006, it was not yet possible for a person to sign on to the Personal Health Record all online, there was a security measure in place that required a person to register online and then receive in the U.S. mail a password to get access. This has since been changed, by the way, it is possible to fully verify for access online. This is useful for this study, because a group of patients did the first step, which can be used for general computer use, and a group of patients did both steps (got the password in the mail, logged back on to established access).

What the table shows is that for patients who did the first step, there was less of an association with their education level or income level with getting access. There is an association with low health literacy with getting access whether or not the patients did the first step registration. In other words, health literacy hurdles followed patients through the two step process, education and income did not. Again, the results are complicated, I’ll try to summarize below.


This graph depicts Kaiser Permanente patients with diabetes AND low health literacy who did the first step registration process and compares the use of certain features to their higher health literacy colleagues. As in the access situation, education and income  did not appear to be a hurdle in actually using PHR functions (that data isn’t pictured). And again, the results are complex, and I have a question as to why the comparison wasn’t between people who had done both steps, since this is required to use these functions. That would have changed the N from 5671 to 4407.

In both situations, Race/Ethnicity was associated with access and use, with less for people identifying as African American, and Latino, more for Asian, less for people identifying as Fliipino.

Glass half-full / glass half-empty (guess which one I am)

At some level these studies can be hard to interpret and put to use to help people because the frame can be different. The authors allude to the idea that this data represents a problem for a health care system that thinks of the Web as way to reduce access to other forms of communication (in person, telephone).

I think if this what a health system is using personal health records for that this is a problem. But this is not what people like me see as the use of personal health records. We see them as access enhancers and ways to model future health systems that don’t yet exist – ones that educate and inform all patients AND their families, allow them to be involved in their care to make it more accurate, of higher quality, while closing gaps within care teams (with team defined as including patient and family, of course).

The way the data is expressed, in the negative, rather than in the positive (“likelihood that someone didn’t use something” is harder to understand than “likelihood that someone did use something”) also may make this hard to visualize.

Therefore, in the glass half-full (or really, 3/4 full) way, this is an excellent piece of work that allows personal health record designers to make health care information more understandable out of the bounds of the exam room conversation. With the advent of mobile technology, specifically, the hurdles presented by technology access will become less so, leaving health literacy as something worth thinking about because it will impact the entire health care experience, not just the online one.

I want to close with one caution that I repeat often – concern about poor usability or lack of health literacy should not be used as a rationale to prevent medical record transparency.

I specifically asked Regina Holliday ( @ReginaHolliday ) about this on our nighttime walk, after I read this study. She told me, “Ted, the medical record is the Bible. We (patients and their families) need to have access to the actual data.”

Even the most unusable personal health record (of which is not, but as an example) with data is better than no data.

This is a paper that I really recommend people reading on their own, so you can get a flavor of the importance of this topic and the amount of effort it took to understand how we can help patients better online. With kudos to the research team and of course the patients in Northern California who provided their experience to the rest of the nation.

I hope to present these at the American Medical Informatics Association Consumer Health Workgroup in Washington, DC, this evening, I’ll let people know if these are clear, or you can let me know in advance in the comments (please!).


Ted Eytan, MD