| Methods |
Study design: double‐blind, cluster, quasi‐randomized placebo‐controlled trial with 2 arms. Alternate families were injected with either iodized oil or saline solution Unit of allocation: family Time period of recruitment and data collection: trial took place in August‐September 1966. Subsequent follow‐ups were carried out in 1967, 1969, 1970 and 1974‐1982 |
| Participants |
Description: entire families including pregnant and non‐pregnant women Location/setting: Jimi River valley in the highlands of Papua New Guinea (at the time accessible only by light aircraft) How participants were recruited and selected: through a census conducted for the purposes of the study in 27 villages Inclusion criteria: none reported Exclusion criteria: none reported Sample size: approximately 16,500 individuals in total (including pregnant and non‐pregnant women) Age: not reported Socioeconomic status: subsistence economy Nutritional status: not reported Baseline iodine status of participants/area: severe deficiency (the area had high incidence of endemic cretinism and high prevalence of goitre). 20% of participants had goitre at baseline. Baseline coverage of iodized salt: not reported Breastfeeding practices: not reported Other important information about participants or setting: none |
| Interventions |
Experimental intervention: injected iodized oil
Total number randomised: n = not reported Timing: before and during pregnancy. Included pregnant women (pregnancy was self‐reported) and women who later became pregnant Form: injected iodized oil Frequency and duration: single dose Iodine compound: not reported Iodine dose: ˜1600 mg iodine (4 mL) oil if 12 years of age or older, ˜800 mg iodine (2 Ll) oil if under 12 years of age (approximately 400 mg elemental iodine per mL) Other nutrients: none Adherence: not reported
Control/comparison intervention: placebo
Co‐intervention(s): none reported |
| Outcomes |
Pregnancy: pregnancy total T4 and TBG assessed in 1970 and 1971 Postpartum: serum‐protein‐bound‐iodine in a subsample of mothers of children with cretinism in 1970 Infant/child: child motor milestones, hearing and speech and squint were assessed from 1967‐1971. Endemic cretinism was diagnosed initially based on motor retardation with deafness and/or squint (though later follow‐ups used different definition). Motor performance was assessed in 1978 and cognitive and motor performance from 1974‐1982. Cretinism (definite versus possible) was assessed again later in a sub‐sample. Since 1972, follow‐ups took place in 5 of the original villages for logistical reasons. Births and deaths were also recorded. Unable to extract infant mortality because it was not reported separately |
| Notes |
We did not adjust for clustering because it was not possible to know if there was more than 1 pregnant woman per family and therefore any clustering effect After it became clear that iodine was an effective prophylactic for endemic cretinism, iodinated oil injections were given to all women of child‐bearing age in each of the villages in 1972. Different reports provide different numbers for cretinism. We used numbers from the Pharoah 1971 report as it provided both the number of children with diagnosed cretinism and the number of children examined and included more villages than later assessments (though the assessments could have taken place beyond 2 years of age, cretinism is a permanent condition so would not have changed after that age) In 6/7 cases of cretinism in the iodine group and 5/26 in the control group, the mothers received the injection during pregnancy |
| Risk of bias |
| Bias |
Authors' judgement |
Support for judgement |
| Random sequence generation (selection bias) |
High risk |
Quasi‐randomized allocation (alternate families) |
| Allocation concealment (selection bias) |
High risk |
Quasi‐randomized allocation (alternate families) |
| Blinding of participants and personnel (performance bias)
All outcomes |
Unclear risk |
Insufficient information to make judgement. Described as double‐blind. Mothers reportedly did not know which group they belonged to though no report specified anything about study personnel |
| Blinding of outcome assessment (detection bias)
All outcomes |
Low risk |
The initial follow‐up assessments and diagnosis of cretinism were made without knowledge of whether the mother received iodine or saline. Cretinism is the only trial outcome included in this review |
| Incomplete outcome data (attrition bias)
All outcomes |
High risk |
Due to remoteness of the region, only 13 or 16 of the original 27 villages were visited for the initial follow‐up visits where cretinism was originally assessed (different numbers provided in different reports). Later follow‐up assessments were only performed in 5 villages for logistical reasons |
| Selective reporting (reporting bias) |
Unclear risk |
Insufficient information to make judgement |
| Other bias |
High risk |
This trial recruited some women (and provided the intervention to these women) before they became pregnant. The intervention may have affected their probability of becoming pregnant, potentially leading to differences between groups and introducing bias |
