Table 4.
Latent class analysis identifying population characteristic patterns among public participants and primary healthcare workers.
| Population | Latent Class | Class Proportion (%) | Key Knowledge Indicator | Key Behavior Indicator | Demographic / Professional Characteristics | Model Fit Indices | Classification Quality | Between-Class Comparison |
|---|---|---|---|---|---|---|---|---|
| Public | 3-class model (optimal) | – | – | – | – | AIC = 21483.6 | Mean posterior probability = 0.83 | – |
| BIC = 21794.2 | ||||||||
| aBIC = 21601.9 | ||||||||
| Low awareness – low screening | 30% | CKD knowledge accuracy: 30% | Screening rate: 19.6% | Age ≥60 years: 45% | – | – | χ2 = 284.7, P < 0.001 | |
| ≤Junior high education: 40% | ||||||||
| Moderate awareness – moderate screening | 45% | CKD knowledge accuracy: 50% | Screening rate: 33.8% | Middle-aged predominant; mixed education levels | – | – | – | |
| High awareness – high screening | 25% | CKD knowledge accuracy: 80% | Screening rate: 80% | ≥High school education: 68.1% | – | – | – | |
| Healthcare Workers | 3-class model (optimal) | – | – | – | – | AIC = 17842.1 | Mean posterior probability = 0.85 | – |
| BIC = 18196.7 | ||||||||
| aBIC = 17984.3 | ||||||||
| Low knowledge – low management capacity | 20% | Knowledge attainment: 30% | Proactive screening: 10% | Shorter working years; | – | – | χ2 = 319.6, P < 0.001 | |
| Regular follow-up: 15% | Public health positions | |||||||
| Moderate knowledge – moderate management capacity | 50% | Knowledge attainment: 50% | Proactive screening: 30% | Mixed professional background | – | – | – | |
| Regular follow-up: 40% | ||||||||
| High knowledge – high management capacity | 30% | Knowledge attainment: 80% | Proactive screening: 70% | ≥10 working years: 60% | – | – | – | |
| Regular follow-up: 75% | Clinical posts: 80% | |||||||
| Adequate laboratory facilities: 70% |
The three-class solutions for both public participants and healthcare workers were selected based on lowest AIC, BIC, and adjusted BIC values. Mean posterior probability: public 0.83, healthcare workers 0.85, indicating acceptable classification quality. Exploratory analysis only: the identified latent classes describe patterns in cognition and behavior but cannot be interpreted as causal. Differences across classes are descriptive and intended to highlight heterogeneity in awareness, knowledge, and management capacity among the study populations.