Cell-Phone Program Improves Diabetes Outcomes, Saves Costs

Miriam E. Tucker

February 03, 2014

A primary-care–based mobile health (mHealth) program that sends health-behavior–related text messages to both type 1 and type 2 diabetes patients both improved outcomes and reduced costs, early data indicate.

The research, conducted within the employee health plan at the University of Chicago, is "the first to demonstrate in a real-world setting value for the patient and value for the health system at large with a mobile-phone–based system," lead author Shantanu Nundy, MD, managing director for program innovations at Evolent Health, Arlington, Virginia, told Medscape Medical News in an interview.

The study, which was published in the February issue of Health Affairs, was conducted while Dr. Nundy was a staff physician at the University of Chicago Medical Center.

The mHealth system isn't just another app, he said. Rather, it engages patients via their own phones with their primary-care providers on a daily basis, sending reminders and reinforcing self-care behaviors. After 6 months, users experienced significant improvements in glycemic control and satisfaction, while the entire health system saw a net 8.8% cost saving.

Such approaches work best in systems where financial incentives are aligned and where the technology can synch with electronic health records (EHRs), neither of which is the norm as yet, Dr. Nundy explained. But, the study does illustrate that "these tools, when built with a behavior change model and when integrated into primary care, can make a difference for patients," he said.

75% of Participants Were Satisfied With Program

The study included a total 348 adult health plan members with diabetes (both type 1 and type 2), including 67 who completed 6 months of enrollment in CareSmarts, an mHealth diabetes program that sends patients automated text messages about self-management throughout the day, on a schedule determined individually at the time of enrollment through a phone conversation with a nurse. The nonparticipants served as controls.

The participants received and replied to an average of 3.4 text messages per day (range, 2.1–6.5 messages). Some messages were reminders such as "Time to check your blood sugar," and some were questions such as "Do you need any refills of your medications?"

Others were educational tips, on topics such as nutrition, glucose monitoring, meds, foot care, and exercise. The patients had all received usual diabetes care, and the messages would often remind them of something they'd learned previously, Dr. Nundy explained.

Alerts are also built into the system to trigger follow-up by a nurse. "Low-level" alerts relate to care coordination issues such as medication refills, which the nurse then coordinates with the primary-care team to resolve.

More serious "high-level" alerts include not responding to the messages or self-reported poor adherence. Here, the nurse calls the patient, assesses their situation over the phone, and emails the results to the primary-care team.

The treatment participants were an average age of 53 years old and had a diabetes duration of 8 years. Two-thirds were African American. About one-third each had HbA1c levels of 7% or less, 7% to 8%, and 8% or greater.

Over the 6 months, participants received an average of 6 phone calls from a nurse, and primary-care providers received an average of 2 communications from a nurse per participant.

Three-quarters of the treatment group said they were satisfied with the program and that they'd be willing to participate in a similar initiative again

Agreement that the program helped them with self-care varied from 59% for medication adherence to 77% for foot care. Most also agreed that the nurse phone calls were helpful, and 88% said that knowing a health professional was reviewing their messages was also important to keeping them engaged.

Aim Is to Get Patients to Help Themselves

Several self-care behaviors improved significantly over the 6-month period, including self-reported healthful eating (from 4.5 days to 5.2 days per week, P = .03), and the number of days spent monitoring their blood glucose (from 4.3 to 4.9, P = .03) or practicing foot care (from 3.6 to 4.3, P = 0.01).

Whether or not the participants' answers are entirely accurate isn't as important as the fact that the questions get them thinking, Dr. Nundy pointed out. "It's really about helping yourself… We're trying to build their own strength and capability to monitor and manage their own condition."

Adherence to medications, measured by the proportion of days covered, increased from 83% to 91% (P = 0.003).

HbA1c levels dropped from 7.9% pretreatment to 7.2% at 6 months posttreatment (P = .01). Among those most poorly controlled (>8%) at baseline, HbA1c levels fell from an average 10.3% to 8.5% (P = .01). In contrast, HbA1c did not change in the controls.

In response to the question, "How satisfied are you with the overall care you receive in your health plan?" with possible answers ranging from 1 ("very satisfied) to 5 ("unsatisfied"), scores improved only in the treatment group, from 1.6 to 1.3 (P = .04).

While the number of diabetes-treatment–related visits didn't change, the number of visits to nondiabetes specialists did, from 3.5 visits in the 6 months prior to the treatment to 2.2 visits 6 months after (P = .007).

Although the reason for this isn't entirely clear, it may relate to a strengthening of the relationship between the patient and the primary-care team. "So, issues were more comfortably managed in primary care than requiring referrals," Dr. Nundy speculated.

Cost Savings and Policy Implications

Total healthcare costs declined by a significant $812 per patient over the 6 months, including a drop of $1332 for outpatient visits along with a $520 increase in prescription drug costs (P = .007 despite the offset).

Costs of the mHealth program were estimated to be $375/participant, suggesting a net cost savings of $437/participant and an 8.8% saving compared with the pretreatment period.

No significant cost differences were seen in the control group. Nonetheless, costs for the entire study population declined significantly (P = 0.02), even though only 20% participated in the program, Dr. Nundy and colleagues point out.

Two current barriers to the success of this type of integrated mHealth system may improve with time, they suggest. For one, it requires aligned incentives, which in the case of the University of Chicago system were present: both the medical center and the employee health plan share a joint mission to improve care while reducing costs.

But that's not the case in the majority of other settings, where stand-alone fee-for-service medical centers would face economic loss with reduced utilization, the authors point out.

However, the growing movement toward accountable-care organizations, in which payers and providers share responsibility for the health of a population and share in the cost savings, will be important for integrated mHealth programs to succeed, Dr. Nundy and colleagues write. Currently, only about 5% to 10% of the US population is enrolled in such integrated health systems, he said.

The other impediment is technological, as the mHealth program isn't currently integrated with the health plan's EHR system, thereby requiring either manual entry or separate archiving of the text-message data. Future "meaningful-use" criteria should include the ability to add mHealth data to patients' EHRs, he and his colleagues say.

And the 2 types of technology shouldn't be confused, Dr. Nundy stressed: "This technology isn't the same as EHR. This is designed to help patients self-manage their conditions."

As the healthcare landscape shifts, he advised healthcare providers to "be on the lookout for opportunities to engage patients in this new and changing environment."

The study was partially funded by the Alliance to Reduce Disparities in Diabetes of the Merck Foundation. Dr. Nundy was supported by the Agency for Healthcare Research and Quality's Health Services Research Training Program. Contributions also came from the University of Chicago Health plan and the Chicago Center for Diabetes Translation Research. Dr. Nundy is a cofounder of mHealth Solutions, which provided the software for the study, but he has no current financial relationship with them. Another cofounder, now full owner, is a study coauthor. Disclosures for the other coauthors are listed in the article.

Health Affairs.2014;33:265-272. Abstract


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.