Does sequencing matter? A cross-over randomized trial to evaluate a bi-level community-based intervention to improve oral hygiene among vulnerable adults

Objectives: To evaluate whether the cumulative effect of an individual-level intervention followed by a building-level intervention, both based on Fishbein's Integrated Model of Behavior Change, has a better effect on oral hygiene clinical outcomes than the sequence of the building-level intervention followed by the individual-level intervention; to determine the added effect of each intervention on the other; to identify the psychosocial mechanisms that might explain the differences.

Methods: Six low-income senior housing complexes were enrolled in the study, and participants were recruited from these buildings. Buildings were randomly assigned to receive either the individual counselling intervention first followed by the building-level intervention, or the building-level intervention first followed by the individual intervention. Participants were assessed for gingival inflammation using the gingival index (GI) and plaque scores (PS) at T0, prior to the interventions, T1, about one month after each intervention and T2, about a month after the interventions switched and were completed. Data were collected on background moderators and cognitive/emotional/behavioural mediators in surveys administered at T0, T1 and T2. General linear mixed models were used to assess changes over time by condition and to analyse the effects of moderators and mediators over time.

Results: Three hundred and thirty-one people completed T0; 306 completed T1 assessments (92.4% retention rate) and 285 completed T2 assessments (86.1% retention rate). All participants improved on GI and PS at T1 and T2 compared to T0. Those in the individual-level intervention condition improved more than those in the building-level condition. Those who were in the building-level intervention followed by the individual intervention continued to improve on GI from T1 to T2. Those in the individual-based intervention followed by the building intervention did not improve significantly from T1 to T2 but remained about the same. For PS, neither group improved significantly from T1 to T2. Several cognitive/behavioural variables significantly affected improvements in GI and PS.

Conclusions: Both interventions were successful in improving GI and PS. The building-level intervention did not provide much additional benefit when it followed the individual intervention although it may have had a sustaining effect. The findings on the cognitive/emotional/behavioural variables support the importance of these factors and should be considered when implementing oral hygiene interventions.