Hoogstraten 1983

Methods	A randomised controlled trial comparing two intervention groups (instruction only and a dental health education film plus instruction) and a control (no instruction).
Participants	Population: 108 adults who had recently registered as a patient in the group practice. Setting: Group practice in Abcoude, Netherlands. Age: 15‐60 years old. Inclusion criteria: n/a. Exclusion criteria: Full dentures, no more than one person per family was admitted to the sample.
Interventions	Two intervention groups and one control: Group 1 (n = 36): Instruction concerning the relationship between sugar consumption and dental health, oral hygiene, the use of fluoride, information about regular visits to the dentist. While presenting the information to the patient, the hygienist performed regular preventive care, such as scaling and polishing. Group 2 (n = 36): Identical standard information as in group 1 with additional film ‐ A Dutch version of 'Four Tons of Teeth' ‐ shown before the same instruction as carried out for Group 1; and presenting more or less the same issues as the instruction. Control group (n = 36): No instruction. Duration: Standard information took 30 minutes; film in group 2 took extra 10 minutes. Personnel conducting interventions: Four dental hygienists.
Outcomes	Outcome: Sugar consumption measured on a scale 1‐5, with a higher score denoting more 'positive' behaviour. Not stated whether this was frequency, timing or amount. Baseline measurement only collected for half of the participants and 6‐12 months after the intervention for all participants. Time points measured: Baseline and 6‐12 months later.
Notes	The dietary intervention was one of a number of interventions delivered together. Both experimental groups received information on oral hygiene as well as diet. Outcome measures included measures of oral hygiene behaviour. Analysis compares baseline and follow‐up rather than intervention and control.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	..."assigned at random" but no further details.
Allocation concealment (selection bias)	Unclear risk	Not reported.
Blinding (performance bias and detection bias) primary outcomes	Unclear risk	Not reported.
Blinding (performance bias and detection bias) secondary outcomes	Low risk	N/A.
Incomplete outcome data (attrition bias) All outcomes	High risk	The authors note that "as usual in longitudinal studies there was a certain drop‐out of subjects"; 150 participants originally recruited (50 per group). 14 participants (28%) dropped out from each group; reasons not provided by group. Intention‐to‐treat analysis not mentioned.
Selective reporting (reporting bias)	Low risk	All important expected outcomes reported.