The protocol for this trial and supporting CONSORT checklist are available as supporting information; see Checklist S1
and Protocol S1
Participants were recruited via local newspaper and radio advertising. Families responding to advertising were sent a Participant Information Booklet (PIB) by post. Those who then wished to enrol telephoned again to request a home visit, usually within 24 hours. Trained investigators used a standard protocol to examine participants for head lice using a plastic detection comb (“PDC”, KSL Consulting, Denmark). If lice were found, and the participant was eligible, a signed consent and assent procedure was followed. Other household members were offered examination and invited to join if eligible.
All enrolled participants provided baseline data on age, gender, hair characteristics, and previous pediculicide use. The lower age limit was 6 months in conformity with the licence for both products; there was no upper limit. Treatments and assessments were conducted in the home. No payment was offered for participation. Ineligible infested household members were provided with 4% dimeticone lotion.
Participants were required to confirm their availability for the duration of the study (14 days following the first treatment) in order to be included in the study. Exclusion criteria were:
- Known sensitivity to any ingredients in the treatments.
- Secondary bacterial infection of the scalp (e.g. impetigo) or any long term scalp condition other than head louse infestation (e.g. psoriasis of the scalp).
- Use of other head louse products within the previous two weeks.
- Use of hair bleach, colour, or permanent wave products, within the previous four weeks.
- Treatment with the antibiotics Co-Trimoxazole or Trimethoprim within the previous four weeks, or taking such a course at the time of enrolment.
- Pregnant or nursing mothers.
- Participation in another clinical study within 1 month before entry to this study.
- Previous participation in this clinical study. 
Prospective participants who wished to participate reported that they understood the purpose and requirements of the study outlined in the PIB and provided written consent. Parents or guardians provided written consent for children below 16 years, who also gave written or verbal assent witnessed by the parent/guardian. Ethical approval was granted by Hertfordshire 1 Research Ethics Committee (EudraCT 2006-004136-73).
The study was conducted in conformity with the principles of the Declaration of Helsinki and of European Union Directive 2001/20/EC.
Dimeticone 4% lotion was supplied in 150 ml bottles (Hedrin® 4% lotion, Thornton & Ross Ltd, Huddersfield, UK) and 0.5% malathion liquid in 200 ml bottles (Derbac-M liquid, SSL International, Manchester, UK). Both products were applied to dry hair, using enough to thoroughly moisten the hair and scalp. Investigators applied the products evenly through the hair using their fingers. Treatments were applied to the full hair length and left to dry naturally. 
The same regimen was repeated 7 days later.
Participants were provided with non-medicated, conditioner free shampoo to ensure all treatments were washed off using the same preparation. Carers were advised of the earliest time treatment should be removed, usually the following morning, and asked not to use louse combs, other form of head louse treatment during participation, and not to divulge the treatment to assessors to maintain blinding. Most participants had previously used one or both preparations so it was impossible to blind them to treatment. However, when asked about the most recent previous treatment it was found only five had used a malathion product, between 2 months and 3 years previously, four of whom were allocated dimeticone and one malathion. Compliance with the protocol was assessed by retrospective questionnaire at each assessment.
The study was designed to compare the efficacy of 4% dimeticone lotion with 0.5% malathion liquid with sufficient power to be able to determine if activity against head lice of either product was superior to the other.
The primary outcome measure was elimination of head lice using two applications of treatment. All participants were examined by dry detection combing, using the “PDC” comb, on days 2, 6, 9, and 14 after the first application of treatment unless they were lost to follow up. Examinations were performed using the comb systematically across the whole scalp. Examinations on days 2, 6, and 9 were limited to 2–3 strokes of the comb on each section, intended to provide diagnostic snapshot data of the status of infestation, because more prolonged combing could have become an additional intervention. A more extensive examination was made on day 14 to try to ensure no lice were present. “Cure” was defined as no lice after the second application of treatment, on days 9 and 14.
Previous experience showed a high risk of reinfestation after cure. 
Knowledge of family circumstances helped identify some reinfestation risks but for statistical purposes we arbitrarily specified criteria for reinfestation as a) no adult lice or third stage nymphs found after the first treatment; and b) on days 9 or 14, no more than two adult lice or third stage nymphs and no younger nymphs found during combing. We acknowledge these criteria could give false outcomes either way but from use over several studies we believe they address the issue of reinfestation without presenting an unreasonably optimistic view of the product efficacy.
Any participant not fitting the cure or reinfestation after cure criteria was categorised as a treatment failure.
A sample size of 31 per group was estimated to have at least 80% power to detect (with 95% confidence) a difference of 35% between the success rates for 4% dimeticone lotion and 0.5% malathion liquid, based on a 70% success rate for dimeticone 4% lotion and evidence suggesting lower success rates with 0.5% malathion liquid, of about 19%–35%. 
The planned sample sizes of 34 per group made some allowance for drop out.
Treatments were randomised using a computer generated list in balanced blocks of 10. Allocation was by inclusion of instruction sheets in numbered sealed envelopes issued in batches of ten to each investigator. A duplicate set was made in the event individual code breaking was required. At enrolment treatment was allocated using the next available number held by the investigator. As randomisation was by individual, household members could receive different treatments. In the event, 73 participants were treated. After the completion of the study, an administrative error had occurred whereby the wrong treatment instructions were included in some of the envelopes. This meant that seven participants originally scheduled to have 0.5% malathion in the randomisation scheme were actually allocated dimeticone 4%. This was discovered during analysis when it was found the individually numbered bottles allocated to some participants did not match the treatment group expected from the randomisation schedule. We knew what treatment a participant had received because their study number and initials were written on the bottle label by the investigator giving treatment. The result was 43 participants were given 4% dimeticone and 30 participants were given 0.5% malathion. Because this error did not compromise blinding of either treatment allocation by investigators in the field, or the assessors assigned to perform the checkups, the viability of the study was not considered to have been impaired, particularly as such a distribution could have occurred naturally as a result of some investigators using only part of their allocation of numbered envelopes. The power to detect a 35% difference with these group sizes was actually very similar to that for the original design.
This study was single blinded because the physical forms of the products are sufficiently different for double blinding to be impractical. Most participants had used one or both preparations previously so it was impossible to blind them to the treatment being used. Different investigators from those applying treatment, blinded to the allocation, performed assessments using “PDC” louse detection combs. Lice found during assessments were removed and fixed to the case record using clear tape. These were later examined under a microscope by another investigator, also blinded to treatment, to determine their developmental stage and if mature, their gender.
For presence/absence variables, Fisher exact tests were used. Differences in success rates between the treatments were quantified by the 95% confidence interval, calculated using a normal approximation to the binomial distribution.
For graded or semi-continuous variables, Kruskal-Wallis analysis of variance was used. As there were only two groups, this was equivalent to using the Mann-Whitney U test.