Table 3.
Association between severe mental illness status and oral health outcomes, NHANES (matched n = 2872), 1999–2016.
| Ordinal Oral Health Outcomes (OR, 95% CI) a | Numeric Oral Health Outcomes (95% CI) | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| DMFT | DT | MT | ||||||||
| Models | Self-Rated Oral Health d | Ache in Mouth e | Tooth Loss f | Periodontal Disease Severity g | Zero-Inflated Model, OR | Count Model, RR | Zero-Inflated Model, OR | Count Model, RR | Zero-Inflated Model, OR | Count Model, RR |
| Sample size n | 2514 | 1533 | 2872 | 1302 | 2872 | 2872 | 2872 | |||
| Unadjusted | 1.84 *** (1.58–2.16) | 1.75 *** (1.43–2.15) | 1.93 *** (1.69–2.21) | 1.11 (0.89–1.39) | 0.89 (0.70–1.13) | 1.28 *** (1.18–1.38) | 1.45 *** (1.17–1.78) | 1.02 (0.93–1.12) | 0.49 *** (0.33–0.71) | 1.37 *** (1.17–1.60) |
| Adjusted for b,c | ||||||||||
| +demographics | 1.61 *** (1.37–1.90) | 1.53 *** (1.24–1.88) | 1.69 *** (1.46–1.94) | 1.02 (0.81–1.29) | 0.97 (0.77–1.23) | 1.22 *** (1.15–1.29) | 1.27 * (1.03–1.56) | 1.05 (0.96–1.14) | 0.64 (0.38–1.05) | 1.41 *** (1.23–1.62) |
| +lifestyles | 1.19 (0.97–1.47) | 1.46 ** (1.14–1.87) | 1.66 *** (1.39–1.99) | 0.77 (0.59–1.01) | 1.08 (0.74–1.58) | 1.16 *** (1.08–1.26) | 1.60 *** (1.19–2.15) | 1.02 (0.93–1.13) | 0.99 (0.59–1.67) | 1.39 *** (1.18–1.64) |
| +comorbidities | 1.18 (0.96–1.45) | 1.43 ** (1.12–1.83) | 1.63 *** (1.36–1.95) | 0.77 (0.58–1.01) | 1.09 (0.74–1.60) | 1.16 ** (1.07–1.25) | 1.57 ** (1.17–2.11) | 1.03 (0.93–1.13) | 1.01 (0.60–1.71) | 1.37 ** (1.16–1.61) |
| +oral hygiene behaviour | 1.19 (0.97–1.47) | 1.42 ** (1.11–1.82) | 1.60 *** (1.34–1.92) | 0.76 (0.58–1.01) | 1.10 (0.75–1.61) | 1.14 * (1.05–1.23) | 1.54 (1.14–2.07) | 1.02 (0.92–1.12) | 0.99 (0.59–1.65) | 1.30 ** (1.10–1.52) |
a OR estimates and 95% CI were pooled over the 5 imputed datasets. b logistic regression or ordinal logistic regression models were performed for ordinal or binary oral health outcomes to assess the inequality of oral health outcomes comparing people with and without SMI. Models are incrementally adjusted for demographics (age, sex, ethnicity, marital status, income), lifestyles (BMI, smoking status, substance misuse, physical activity, sugar intake), comorbidities (diabetes, cardiovascular disease), and oral health behaviour (dental visit, dental floss use). c Zero-inflated negative binomial (ZINB) models were performed to assess the inequality of dental caries experience comparing people with and without SMI. ZINB model is a 2-part model, with zero-inflated model predicting the chance of excessive zeros, and the count model predicts the number of DMFT/DT/MT. For example, in DMFT unadjusted model, zero-inflated odds ratio 0.94 indicated that people with SMI has 6% (1–0.94) lower chance of being caries-free (DMFT = 0), and the count model RR as 1.17 showed that people with SMI have 17% higher chance to have more dental caries than people without SMI. d self-rated oral health: excellent to good (ref), fair to poor. e Mouthache: never or hardly ever (ref), occasionally to very often. f tooth loss: no loss (ref), 1–10, 11–20, 21–31, and edentulous. g periodontal disease severity: none (ref), mild to moderate, severe. OR, odds ratio; RR, rate ratio; DMFT, number of decayed, missing and filled teeth; DT, number of decayed teeth; MT, number of missing teeth due to decay. * p < 0.05, ** p < 0.01, *** p < 0.001.