Assessing Community Readiness for Preventing Youth Substance Use in Colombia: A Latent Profile Analysis

Patterns of Sustainability Capacity Among Organizations That Deliver the National Diabetes Prevention Program: A Latent Profile Analysis

INTRODUCTION

Since the launch of the National Diabetes Prevention Program (DPP) in 2010, more than 3,000 organizations have registered with the Centers for Disease and Control and Prevention to deliver the program; today, however, only approximately 2,000 organizations are registered, indicating challenges with sustainability. We used the Program Sustainability Assessment Tool (PSAT) to explore patterns of sustainability capacity among National DPP delivery organizations.

METHODS

We used data from a cross-sectional online survey conducted in August and September 2021 of staff members (N = 440) at National DPP delivery organizations. We conducted a latent profile analysis to identify latent subpopulations on the basis of respondent PSAT domain scores. Regression analyses were used to estimate associations between derived latent classes, PSAT scores, and respondent characteristics.

RESULTS

The 4-class model included 4 groups of capacity for program sustainability, ranging from low to high: low (class 1) with 8.0% of the sample, medium-low (class 2) with 22.0%, medium-high (class 3) with 41.6%, and high (class 4) with 28.4%. Program evaluation (mean score = 5.1 [SD = 1.4]) and adaptation (mean score = 5.3 [SD = 1.3]) were the domains with the highest scores, while funding stability (mean score = 4.0 [SD = 1.6]) and Partnerships (mean score = 4.0 [SD = 1.7]) had the lowest scores. In our sample of National DPP delivery organizations, most reported relatively high capacity for program sustainability, and key indicators associated with sustainability capacity were virtual delivery, location of delivery, funding sources, and organization type.

DISCUSSION

Similar to sustainability capacity findings from other PSAT studies, our study found that funding stability and partnerships are areas to strengthen. This insight is useful in sustainability planning at organizational and national levels across multiple programs.

Robustness of Latent Profile Analysis to Measurement Noninvariance Between Profiles

Latent profile analysis (LPA) identifies heterogeneous subgroups based on continuous indicators that represent different dimensions. It is a common practice to measure each dimension using items, create composite or factor scores for each dimension, and use these scores as indicators of profiles in LPA. In this case, measurement models for dimensions are not included and potential noninvariance across latent profiles is not modeled in LPA. This simulation study examined the robustness of LPA in terms of class enumeration and parameter recovery when the noninvariance was unmodeled by using composite or factor scores as profile indicators. Results showed that correct class enumeration rates of LPA were relatively high with small degree of noninvariance, large class separation, large sample size, and equal proportions. Severe bias in profile indicator mean difference was observed with intercept and loading noninvariance, respectively. Implications for applied researchers are discussed.