Improving Diabetic Patient Health Through Assistive-Reading Technology

Diabetes self-management requires knowledgeable communication and capacity for self-management. People with type 2 diabetes mellitus (T2DM) must learn from health care providers about the condition to build knowledge on how to manage the disease and competently carry out self-management. T2DM self-management is extensive: estimates of time spent on recommended self-management activities can exceed 10 hours per week, equivalent to part-time employment. Effective T2DM-related self-management requires patients to read, understand, and execute care plans. In short, T2DM comes with a demanding workload that assumes patients have enough capacity to read and follow complex instructions to successfully manage their disease.

Barriers to building understanding and competence for self-management both reflect and contribute significantly to inequality. Existing social systems in the United States have provided inequitable educational achievement, including literacy in general and health literacy in particular. The result is that the U.S. has widespread and entrenched adult illiteracy. 18% of U.S. adults (roughly 44 million) are unable to decipher unfamiliar words or identify information in short texts. In practical terms, a large segment of the U.S. population is without the skills to read medicine or nutrition labels, follow written aftercare instructions, etc. This deficit is reflected in the T2DM patient population. Large percentages (e.g., 35-55%) of individuals with T2DM have reading levels at 6th grade or below, with lower literacy significantly associated with several demographic variables (e.g., lower socioeconomic status), indicating inequitable distribution. Moreover, lower functional English and health literacy, and poorer diabetes knowledge, are associated with poorer glycemic control.

However, work on facilitating the entry of adults with limited literacy and English proficiency into the workforce demonstrates that these challenges can be successfully addressed through an integrated adult literacy approach. This approach of "integrated adult literacy" combines the delivery and understanding of concepts and information with in-the-moment assistive-reading technology and contextualized literacy instruction regardless of whether an adult employee is correctly perceived to be at risk for missing critical information.

Health care supports for capacity in self-management lack appreciation for integrated adult literacy and have limited effectiveness. Health literacy is not only an individual issue, but also reflects if systems and materials are calibrated to individual capacity. Yet health care does not typically operate in recognition of this fact. Written patient materials for post-visit care are consistently complex, often far exceeding the intended consuming audience's reading level. Meanwhile, practices such as "Teach Back" and including interpreters in clinic visits have short-term benefits for patient understanding and ability to carry out self-management; as these effects diminish over time, mastery of key information becomes elusive. The inadequacy of these measures is compounded by stigma that can accompany low literacy or limited English proficiency. Even patients who have developed literacy skills may be embarrassed about a lack of understanding. Stigma, embarrassment, or simply the stress of a clinical appointment can prevent patients from asking clarifying or follow-up questions, a process integral to establishing content mastery.

Existing systems are inadequate to bridging gaps in patients' needs and readiness to manage diabetes. Achieving universal literacy in the U.S. adult population historically has been a major challenge; systems in place lack sufficient system capacity and scale to address it. For example, only a fraction of U.S. adults in need of reading instruction have access to it. Literature indicates a growing recognition by health educators of the potential for engaging with and implementing techniques from adult basic education, including integrated adult literacy. However, significant factors would still limit the impact which implementing such measures could have for patients with limited literacy or English proficiency. Foremost would be the significant issue of scale needed for impact and improvement.

A paradigm shift is needed, from clinic-by-clinic intervention to highly scalable, low cost in-the-moment support using assistive technology (AT) but little is known, particularly from an integrated adult literacy perspective. As noted, existing systems are inadequate to increase integrated adult literacy generally, and clinical interventions require patients to retain knowledge as they use complex, written instructions and education materials in self-management away from the clinic. Because changes to these have historically remained out of reach, the investigators propose shifting the paradigm away from focusing on providers' capacity to communicate and educate, and toward supporting patients' capacity to access, comprehend and internalize health education through AT, expanding their capacity for self-management in the moment, on an as-needed basis. AT commonly refers to any equipment, software, etc. that is used to maintain or improve the functional capabilities of individuals with disabilities, often with features and functions that can be adjusted to an individual's specific requirements. Notably, existing electronic health (eHealth) and mobile health (mHealth) support technologies with text-heavy interfaces have not incorporated reading AT and are frequently limited to single-use applications with poor readability, and/or usability issues, negating their benefit for integrated adult literacy for diabetes self-care.

Substantial evidence describes how AT equalizes information access across education levels. When combined with the near-ubiquity of mobile devices and telecommunications networks, AT - and specifically assistive reading technology - holds the potential for universal, integrated adult literacy supports that could substantially impact diabetes self-management education. Sidewalk cutouts offer a useful analogy for whole-population benefit. While cutouts expand mobility to those using walkers and wheelchairs, they also provide unintended benefits to any child riding a bicycle or parent pushing a stroller. Similarly, benefits from reading AT could extend beyond the intended population (e.g., adults with limited functional literacy or English proficiency) to those with diminished eyesight and even fully literate adults simply struggling with difficult materials.

The current study will use a small pilot trial design of the GogyUp reader app for use with print materials versus no support for reading capacity among adults with diabetes.

Primary Endpoint: three-month (follow-up) scores on two subscales of the Health Literacy Questionnaire: Subscale 9: Understanding health information well enough to know what to do ("Understanding health information"); Subscale 2: Having sufficient information to manage my health ("Having sufficient information").

Secondary Endpoint: three-month (follow-up) score on the Perceived Diabetes Self-Management Scale (PDSMS).

Analysis: We will compare the mean outcomes at follow-up between the treatment arms using analysis of covariance (ANCOVA), which compares follow-up outcome values between study arms while controlling for the baseline levels to gain power to detect an effect. The primary analysis will be an intent-to-treat (ITT); for non-monotonic missing (e.g., missing a baseline value), we will use multiple imputation with pattern mixture model adjustments in sensitivity analyses allowing for plausible differences among the missing-the justification for this is that pattern mixture models allow for a Missing Not At Random assumption, which is plausible in this case because we will not have measures for all likely variables associated with missingness. Pattern mixture models allow for us to set plausible values for missingness at both high or low levels to obtain a range of outcome estimates allowing for plausible missing data scenarios. For any monotonic missing- i.e., attrition (withdrawal or other lost to follow-up), we will use last observation carried forward but will also conduct sensitivity analyses using pattern mixture models under plausible assumptions of missing data patterns. We will present results using adjusted (least-squares) means with standard errors, p-values, and 95% confidence intervals. Baseline levels of endpoint outcomes will be included as a covariate to gain power to detect an effect, a standard use of ANCOVA.

The maximum sample size is up to 200 individuals, with expectation of potential non-contact by patients after a recruitment mailing, screening failures and refusal bringing likely sample size down to approximately 120-160. The target population are individuals aged 18-85 diagnosed with type 2 diabetes mellitus seen at clinics serving large proportions of low-income patients (e.g., safety net clinics). We will attempt to sample up to 200 individuals from the MHealth Fairview health system in the seven-county Twin Cities Metropolitan area.

Study activities will not take place in any clinical sites; participants will be identified based on having been seen for a clinic visit to a primary care clinic within the partnering health system and will receive a mailing from the health system providing contact information for the researchers; patients will themselves contact the research team if they choose. The study will not include sites outside of the United States.

Contribution: Understanding use and potential benefit of in-the-moment reading support for adult patients in understanding diabetes self-management written materials. The potential benefit of reading AT, readily embedded in adult patients' lives, that expands their capacity to use written materials at scale, at low cost, and particularly from an integrated adult literacy standpoint, has not been previously studied.