Online Program

Behavior change as a mediator of the effect of nurse-community health worker team intervention on diabetes control in American Samoa

Sunday, November 3, 2013

Shira Dunsiger, PhD, Centers for Behavioral and Preventive Medicine and the Department of Behavioral and Social Sciences, The Miriam Hospital/Brown University, Providence, RI
Stephen McGarvey, PhD, MPH, Department of Epidemiology, Brown University School of Public Health, Providence, RI
Background: Type 2 diabetes prevalence among adults in American Samoa (AS) is 21.5%, compared to 11% of adults in the U.S. overall. We conducted a primary care-based nurse-community health worker (CHW) team intervention that was adapted for this setting and cultural context. Initial findings showed the CHW group significantly improved diabetes control, compared to usual care at 12-month follow-up. Objective: This paper examines whether the CHW intervention effect on HbA1c was mediated by changing health behaviors and primary care provider (PCP) visits. Methods: 268 type 2 diabetes patients were enrolled from Tafuna Clinic (TC), a federally qualified community health center in AS, and randomized by village to a CHW intervention versus wait-list control. The primary outcome was HbA1c and hypothesized mediators were high fat diet (>35% fat, Samoan food frequency questionnaire), medication adherence (% with no non-adherence behaviors, Hill-Bone measure), at least moderate physical activity (% ≥600 METS, WHO STEPS), PCP visit counts in years prior and during CHW intervention. Using the product of coefficients method with bootstrapped standard errors, we tested the simultaneous effects of the posited mediators and estimated direct and indirect intervention effects. Models adjusted for mediator baseline values, along with potential confounders. Results: Results showed the intervention reduced HbA1c by increasing healthy eating (indirect effect=-0.03, se=0.01) and decreasing medication non-adherence (indirect effect=-0.04, se=0.01). Taken together with the other mediators, physical activity and PCP visits were not significant treatment effect mediators. Conclusions: This nurse-CHW team intervention significantly improved all targeted health behaviors in CHW versus control groups (p's<.05). Model results suggest that medication adherence and healthy eating were key factors in HbA1c change, accounting for 73% of variance. The CHW role can be a valuable addition to the primary care team for at-risk patients in resource-poor settings.

Learning Areas:

Biostatistics, economics
Public health or related research
Social and behavioral sciences

Learning Objectives:
Describe how the use of mediation analysis can help to explore mechanisms of change in intervention research Identify the most beneficial roles for community health workers on a primary care team toward improving diabetes control, based on research findings.

Keyword(s): Diabetes, Statistics

Presenting author's disclosure statement:

Qualified on the content I am responsible for because: I am a biostatistician and have served as a PI or Co-Investigator on numerous NIH-funded research grants, including both methodological and behavioral research. Among my scientific interests has been identifying mediators of treatment effects in behavioral medicine.
Any relevant financial relationships? No

I agree to comply with the American Public Health Association Conflict of Interest and Commercial Support Guidelines, and to disclose to the participants any off-label or experimental uses of a commercial product or service discussed in my presentation.