Study characteristics

Methods

Study design: RCT

Study location/setting: Boston, USA. Urban area

Recruitment: children were screened for lead from May 1993 to April 1995

Follow‐up: 6 months from baseline

Intention‐to‐treat: available‐case analysis

Power calculation: performed to determine number of participants (required number not recruited)

Participants

Eligibility criteria

1. Resided in the city of Boston

2. < 4 years of age

3. Had venous BLL of 11.0–24.0 μg/dL

4. No history of lead poisoning or chelation therapy

5. Not expected to undergo chelation treatment

6. Lived on the premises for ≥ 3 months with no definite plans to move within the next 3 months

7. Lived in home with lead‐based paint on ≥ 2 window sills or window wells (or both), as determined by sodium sulphide tests

8. The home had not been previously de‐leaded or received lead hazard reduction activities

9. The parents spoke English, Spanish or Cape Verdean Creole

10. No other child in the home was already a study participant

Participation rate: 63/402 (16%) enrolled, of whom 41 were randomised (22 intervention, 19 control); 22 other participants at high risk were automatically assigned to the intervention and therefore were not considered in this Cochrane Review

Reason for non‐participation: 163 unreachable; 64 unable to communicate due to language barriers; 112 refused to participate (demographic characteristics similar between participants and non‐participants)

Analysis: 24/41 (59%) for BLLs, 22/41 (54%) for household dust

Number of dropouts/withdrawals: 17 BLLs, 19 household dust lead levels

Reasons for dropout/withdrawal: children were excluded because no 6‐month follow‐up blood samples were taken, their homes received non‐study environmental interventions or they received chelation therapy (no specific numbers per reason were reported for the randomised children that dropped out)

Intervention baseline characteristics (available for n = 11)

1. 36.4% boys, 63.6% girls

2. Mean age 28.3 months

3. BLL 17.6 µg/dL

Control baseline characteristics (available for n = 13)

1. 53.9% boys, 46.1% girls

2. Mean age 23 months

3. BLL 16.3 µg/dL

Interventions

Intervention (low technology lead hazard reduction)

1. Remove lead dust

2. Loose paint chips

3. HEPA vacuum

4. Parental education: demonstrating effective housekeeping techniques and sending monthly reminders with instructions to wash hard surface floors, window sills and wells

5. Usual outreach and educational activities (1 home visit by an outreach worker to visually assess the home for lead hazards and to educate the carer about the causes and prevention of lead poisoning)

Control

1. Usual outreach and educational activities provided to both groups (1 home visit by an outreach worker to visually assess the home for lead hazards and to educate the carer about the causes and prevention of lead poisoning)

Outcomes

1. BLL 6 months from baseline (venous blood sample)

2. Household dust lead levels

Notes

Funding: this research was supported by a co‐operative agreement (Grant H64/CCH108235‐03) with the CDC, Atlanta, GA

Conflicts of interest: none declared

Other comments

1. Different baseline characteristics and small sample size

2. Inconsistent parental compliance with housekeeping

3. Several participants had non‐study interventions and were excluded from analysis in the report

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Low risk

Quote from correspondence with author: "an open list of random numbers."

Allocation concealment (selection bias)

High risk

Quote from correspondence with author: "open list."

Blinding of participants and personnel (performance bias): objective outcomes

Low risk

Quote from correspondence with author: "the subjects and the investigators interacting with the subjects knew which group they were assigned to."

Comment: we rated this domain at low risk of bias because BLLs and household dust lead measures are unlikely to be influenced by participants' knowledge about treatment allocation.

Blinding of outcome assessment (detection bias)
All outcomes

Low risk

Quote from correspondence with author: "lab analysers were blinded."

Incomplete outcome data (attrition bias)
Blood lead level

High risk

Comment: high attrition rate (41%) and high differential attrition. Attrition rate was 18% points higher in the intervention group than in the control group.

Incomplete outcome data (attrition bias)
Household dust lead level

High risk

Comment: high attrition rate (46%) and high differential attrition. Attrition rate was 27% points higher in the intervention group than in the control group.

Selective reporting (reporting bias)

Unclear risk

Comment: insufficient information.

Other bias

Low risk

Comment: study appeared free from other sources of bias.

Study characteristics

Methods

Study design: RCT

Study location/setting: Australia. Mining community

Recruitment: October 1994 to August 1996

Follow‐up: 12 months from baseline

Intention‐to‐treat: available‐case analysis

Power calculation: performed to determine number of participants (required number not recruited)

Participants

Eligibility criteria

1. Children from a surveillance programme, aged 12–60 months

2. BLL 15.0–30.0 μg/dL

Participation rate: 103/365 (28%). 365 were eligible for remediation, 117 enrolled in remediation programme and 103 were eligible for randomisation (all children with a BLL ≥ 30.0 μg/dL were offered immediate home remediation, and therefore were not part of the randomisation). Of the 103 children, 90 were matched by age and BLL range and were randomised (45 intervention, 45 control)

Reason for non‐participation: 13 were unable to be adequately matched

Analysis: 88/90 (98%) blood levels analysed

Number of dropouts/withdrawals: 2

Reasons for withdrawal: no blood samples for either participant

Baseline data available for all children randomised

1. 42% boys, 58% girls overall

2. Mean age 3.5 years

3. Mean BLL 19.4 µg/dL

Baseline characteristics not reported separately for intervention and control group

Interventions

Intervention (home remediation work)

1. Performed on intervention households and varied depending on assessment of need to provide each house with a "similar level of lead safety"

2. Work may have included: ceiling dust removal, sealing of ceilings, paint stabilisation, replacement of floor coverings/windows and cleaning

3. Visiting families at home and providing them with information about minimising lead hazards

Control

1. Visiting families at home and providing them with information about minimising lead hazards

Outcomes

1. BLL 6 months from baseline (venous blood sample)

(Internal floor dust quintile only used to examine dose‐response effects)

Notes

Funding: Australian Government Department of Health and Ageing

Conflicts of interest: none declared

Other comments

1. Control group received remediation after completion of study

2. To examine dose‐response effect, indoor dust levels were measured to examine the extent in which indoor lead levels were associated with changes in BLL

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Low risk

Quote from correspondence with author: "children were matched in pairs and then a coin tossed to see which would be the 'case' and have their home re‐mediated first."

Allocation concealment (selection bias)

Unclear risk

Comment: insufficient information.

Blinding of participants and personnel (performance bias): objective outcomes

Low risk

Comment: no information provided on blinding. However, we rated this domain at low risk of bias because BLLs and household dust lead measures are unlikely to be influenced by participants' knowledge about treatment allocation.

Blinding of outcome assessment (detection bias)
All outcomes

Low risk

Quote from correspondence with author: "lab analysers were blinded."

Incomplete outcome data (attrition bias)
Blood lead level

Low risk

Comment: attrition rate was low (2%).

Selective reporting (reporting bias)

Low risk

Comment: study protocol NA, but the published reports included all expected outcomes, including those that were prespecified and confirmed by investigator.

Other bias

Low risk

Comment: study appeared free from other sources of bias.

Study characteristics

Methods

Study design: RCT (embedded in the HOME cohort study)

Study location/setting: Cincinnati, Ohio, USA; urban, suburban and rural areas

Recruitment: pregnant women attending 1 of 9 antenatal care clinics affiliated with 3 hospitals between 1 March 2003 and 31 January 2006

Follow‐up: at 1, 2, 3, 4, 5 and 8 years of age

Analysis: intention‐to‐treat and per‐protocol

Power calculation: 80% power to detect a 3.4‐point difference in IQ with 180 children per group and 3.4 µg/dL difference in blood lead concentration between the intervention and control groups

Participants

Eligibility criteria for mothers

1. < 19 weeks' pregnant

2. ≥ 18 years of age

3. Residence in a house built in or before 1978

4. Not living in a mobile or trailer home

5. HIV‐negative status

6. Not taking medications for seizures or thyroid disorders

7. Plan to continue antenatal care and to deliver at the participating clinics and hospitals

8. Plan to live in the greater Cincinnati area for the next year

9. English fluency

10. No diagnoses of diabetes, bipolar disorder, schizophrenia or cancer (that requires radiation treatment or chemotherapy)

Participation rate: 355 pregnant women randomised (174 intervention group, 181 control group)

Reason for non‐participation: not interested, dropped out in run‐in period, landlords refused

Analysis: 320/355 (90.1% analysed)

Number of dropouts/withdrawals: 35/355 (9.9%)

Reasons for dropout/withdrawals: not reported

Intervention baseline characteristics

1. 51.1% boys, 48.9% girls (n = 174)

2. Infant gestational age at delivery 39.0 weeks (SD 1.5) (n = 166)

3. Geometric mean pregnancy blood level 0.7 µg/dL (SD 1.5) (n = 167)

Control baseline characteristics

1. 42.5% boys, 57.5% girls (n = 181)

2. Infant gestational age at delivery 39.0 weeks (SD 2.0), (n = 177)

3. Geometric mean pregnancy blood level 0.7 µg/dL (SD 1.5) (n = 171)

Interventions

Intervention

1. By 32 weeks' gestation and before delivery, the intervention group received a combination of interventions to reduce exposure to residential lead hazards: covering bare lead‐contaminated soil with groundcover, installing a tap water filter if the lead concentration in drinking water exceeded 2 μg/L, repairing/repainting peeling or deteriorating lead‐based paint, creating smooth and cleanable floors and windows, installing window trough liners, replacing windows that have lead‐based paint or show more than 10% deterioration, and undertaking extensive dust control and cleanup; additional cleaning, if necessary.

Control

1. Women received injury prevention devices or residential modifications before their children were 6 months of age

Outcomes

1. Dust lead measurements: floor, interior window sill and window troughs in the main activity room, child's bedroom and kitchen were collected approximately at 20 weeks' gestation and when the children were 1 and 2 years of age

2. BLL: collected blood from women at 16 weeks' gestation and 26 weeks' gestation and shortly before, or within, 48 hours of delivery as well as from children at 1, 2, 3, 4, 5 and 8 years of age

3. Neurobehavioural outcomes

   1. Behaviour Assessment System for Children, Second Edition, a parent‐reported questionnaire to assess behavioural problems at 1–8 years of age

   2. Bayley Scales of Infant Development, Second Edition, to assess children's mental and psychomotor development at 1, 2 and 3 years of age

   3. Behaviour Rating Inventory of Executive Function, a parent‐reported questionnaire to measure children's executive functions at 3, 4, 5 and 8 years of age

   4. Conners' Continuous Performance Test to measure children's execution functions at 5 and 8 years of age

   5. Wechsler Preschool and Primary Scale of Intelligence, Third Edition, used at 5 years of age, and the Wechsler Intelligence Scale for Children, Fourth Edition, used at 8 years of age to evaluate children's cognitive abilities

Notes

Funding: supported by grants from the National Institutes of Environmental Health Sciences, the US Environmental Protection Agency, and the US HUD

Conflicts of interest: Dr Lanphear reported serving as an expert witness in childhood lead poisoning cases for which he has not received any compensation. Dr Braun reported being financially compensated for conducting a re‐analysis of a child lead exposure study for the plaintiffs in a public nuisance childhood lead poisoning case. No other disclosures were reported.

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Low risk

Quote from report: "Using random number generation, we assigned the remaining 355 women (75.8%) in blocks of 10" (p. 935).

Allocation concealment (selection bias)

Low risk

Quote from report: "sealed the assignment codes in radio‐opaque envelopes until the research assistants confirmed each participant's eligibility" (p. 935).

Blinding of participants and personnel (performance bias): objective outcomes

Low risk

Comment: according to the protocol (supplement 1), the trial was a nested, randomised, single‐blinded trial. We could not find any additional information on blinding in the protocol. We rated this domain at low risk of bias for objective outcomes because BLLs and household dust lead measures are unlikely to be influenced by participants' knowledge about treatment allocation.

Blinding of participants and personnel (performance bias): subjective outcomes
All outcomes

Unclear risk

Comment: according to the protocol (supplement 1), the trial was a nested, randomised, single‐blind trial. We could not find any additional information on blinding in the protocol. We rated this domain as unclear risk for subjective outcomes, because neurobehavioural outcomes could be influenced by lack of blinding.

Blinding of outcome assessment (detection bias)
All outcomes

Unclear risk

Comment: no information if outcome assessors were blinded. According to the protocol (supplement 1), the trial was a nested, randomised, single‐blind trial. We could not find any additional information on blinding in the protocol, so we rated the risk of detection bias unclear for subjective and objective outcomes.

Incomplete outcome data (attrition bias)
Blood lead level

Low risk

Comment: low attrition rate (9.9%) and low differential attrition (for BLL and neurobehavioural outcomes). Attrition rate was 10.3% in the intervention group and 9.4% in the control group (0.9 percentage points difference).

Incomplete outcome data (attrition bias)
Household dust lead level

Low risk

Comment: low attrition rate (9.9%) and low differential attrition. Attrition rate was 10.3% in the intervention group and 9.4% in the control group (0.9 percentage points difference).

Selective reporting (reporting bias)

Low risk

Comment: trial protocol was registered in ClinicalTrial.gov and provided as supplement 1.

Other bias

Low risk

Comment: study appeared free from other sources of bias.

Study characteristics

Methods

Study design: RCT

Study location/setting: Rhode Island, USA. Urban area

Recruitment: all children who were identified through routine blood lead testing as having venous BLLs of 15.0–19.0 µg/dL, and reported to Rhode Island Department of Health between July 1999 and June 2002, were referred to the study

Follow‐up: 12 months from baseline

Intention‐to‐treat analysis: available‐case analysis

Power calculation: performed to determine number of participants (required number recruited)

Participants

Eligibility criteria

1. Venous BLL 15.0–19.0 μg/dL

2. ≤ 28 months of age

3. Family spoke English or Spanish

Participation rate: 175/241 (73%) consented to participate and were randomised (92 intervention, 83 control)

Reason for non‐participation: 66 refused due to work and school responsibilities (no breakdown in figures per reason reported)

Analysis: 145/175 (83%) analysed BLL, 153/175 floor dust lead levels

Number of dropouts/withdrawals: 30 BLL, 22 floor dust lead levels

Reasons for dropout/withdrawal: 9 children moved away, 2 were lost to follow‐up, 9 parents refused participation of their child during the study, 2 refused first and all subsequent visits. No reason for dropout/withdrawal was reported for 8 children

Intervention baseline characteristics (available for n = 90)

1. Mean age 19.1 months

2. Mean BLL 16.5 µg/dL

Control baseline characteristics (n = 83)

1. Mean age 18.8 months

2. Mean BLL 16.3 µg/dL

Sex of children not reported

Interventions

Intervention

1. Parental education (with nursing care plan) via 5 home visits during 1‐year period. Nurses collected interior dust and soil samples, evaluated parent–child interaction, identified occupational or recreational exposure to lead sources and other factors thought to influence lead exposure. The nursing care plan directed parent teaching and other services

Control

1. Children received customary care: 1 or 2 educational visits by outreach workers. These visits focused on standard health education about lead poisoning and its prevention but did not include environmental sampling, education tailored to individual circumstances or assessment or parent–child interaction

Outcomes

1. BLL 12 months from baseline (venous blood sample)

2. Household floor and window dust lead

(Questionnaires on lead exposures)
(Parental–infant interaction scale)

Notes

Funding: sponsored in part by CDC (grant TS 275 14/14) and Maternal and Child Health Bureau (grant 5T76 MC 00001; formerly MCJ201)

Conflicts of interest: none declared

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Low risk

Quote from report: "[r]andom numbers table was used to assign cases to either the intervention or the comparison group, sequentially" (p. 148).

Allocation concealment (selection bias)

Low risk

Quote from report: "group assignments were sealed into envelopes and unknown to either study personnel or the families until after parental consent was obtained" (p. 148).

Blinding of participants and personnel (performance bias): objective outcomes

Low risk

Quote from report: "the nurses who provided follow up to comparison group children were blinded and nurses that provided care to intervention group were not blinded" (p. 148).

Comment: we rated this domain at low risk of bias because BLLs and household dust lead measures are unlikely to be influenced by participants' knowledge about treatment allocation.

Blinding of outcome assessment (detection bias)
All outcomes

Low risk

Quote from report: "venous blood samples were collected by children's pediatric health care providers" (p. 148).

Incomplete outcome data (attrition bias)
Blood lead level

Low risk

Comment: the overall attrition rate (17%) was acceptable. The attrition rates in intervention (18%) and control group (15%) were similar.

Incomplete outcome data (attrition bias)
Household dust lead level

Low risk

Comment: the overall attrition rate was acceptable (13%). The attrition rates in intervention (12%) and control group (13%) were similar.

Selective reporting (reporting bias)

Low risk

Comment: study protocol NA but it was clear that the published reports included all expected outcomes, including those that were prespecified and confirmed by the investigator.

Other bias

Low risk

Comment: study appeared free from other sources of bias.

Study characteristics

Methods

Study design: RCT

Study location/setting: Philadelphia, USA. Urban area

Recruitment: children were recruited from urban outpatient practices located in low‐income neighbourhoods of Philadelphia

Follow‐up: 24 months from baseline

Intention‐to‐treat analysis: available‐case analysis

Power calculation: performed to determine number of participants (required number recruited)

Participants

Eligibility criteria

1. Residing in Philadelphia county

2. Children speak English or Spanish

3. Home that was judged to be in a condition enabling remediation

4. No history of elevated BLLs

5. No former participation in the Lead Safe Babies programme

6. No child of the family has ever received services from the Childhood Lead Poisoning Prevention Program of the Philadelphia Department of Public Health

Participation rate: 314/314 (100%) newborn children enrolled and randomised (154 intervention, 160 control); 310/310 (100%) households enrolled

Reasons for non‐participation: NA

Analysis: 279/314 (89%) BLLs analysed at 12 months of age; 110/306 (36%) household dust analysed at 12 months. No information on number of children for whom BLLs were analysed at 24 months

Number of dropout/withdrawal: 35 (BLLs), 196 (household dust lead)

Reasons for dropout/withdrawal: for 35 children, no venous specimen was taken (or reasons specified); for 196, no household dust lead level was measured because of problems finding participants who changed address or telephone numbers, non‐compliance with study visits, lack of approval by family members (no breakdown in specific numbers per reasons reported)

Intervention baseline characteristics (n = 154): 53.2% boys, 46.8% girls

Control baseline characteristics (n = 160): 51.2% boys, 48.8% girls

Mean age and mean BLL only reported for both groups combined: mean age: 11 months and mean BLL: 2.7 µg/dL

Interventions

Intervention

1. Standard lead poisoning prevention education plus additional extensive education regarding maintaining home in lead‐safe condition and home visits from outreach workers at baseline, 6 and 12 months. The additional education was compiled in a 22‐page handbook.

2. Cleaning materials provided (MEG)

Control

1. Standard lead poisoning prevention education, not described in more detail (SEG)

Outcomes

1. BLLs at 12 and 24 months (venous blood sample)

2. Housing lead dust levels at 12 months

(Parental Knowledge Assessment)

Notes

Funding: HUD Lead Technical Studies Grant

Conflicts of interest: none declared

Other comments: a matched comparison group was included in results of the paper, receiving community standard for prevention of elevated BLLs. This group was not part of the randomisation process and, therefore, was not integrated in this Cochrane Review

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Low risk

Quote from correspondence with author: "randomised blocks using computer‐generated random numbers."

Allocation concealment (selection bias)

Unclear risk

Quote from correspondence with author: "study coordinator selected next card in the random sequence to randomise that family."

Blinding of participants and personnel (performance bias): objective outcomes

Low risk

Quote from correspondence with author: "once the randomization occurred they were told of their assignment. The outreach workers who performed the randomization were made aware of the assignment category, as well."

Comment: we rated this domain at low risk of bias because BLLs and household dust lead measures are unlikely to be influenced by participants' knowledge about treatment allocation.

Blinding of outcome assessment (detection bias)
All outcomes

Unclear risk

Quote from correspondence with author: "the samples were sent to two different analytic labs during the course of the 3‐year study period, and their job is to analyze BLLs. They were blinded to status". "The outreach workers collecting the dust wipe samples knew the household assignment" – this might have had an influence on household dust levels (comment).

Incomplete outcome data (attrition bias)
Blood lead level

Low risk

Comment: attrition rate for BLL was acceptable (11%). No information on dropout rates in either study arm.

Incomplete outcome data (attrition bias)
Household dust lead level

High risk

Comment: for the outcome, household dust, attrition rate was very high (64%). No information on dropout rates in either study arm.

Selective reporting (reporting bias)

Low risk

Comment: study protocol NA, but the published reports included all expected outcomes, including those that were prespecified and confirmed by the investigator.

Other bias

Low risk

Comment: study appeared free from other sources of bias.

Study characteristics

Methods

Study design: quasi‐RCT (even/odd clinic number assignment)

Study location/setting: Baltimore, USA. Recruited from a lead poisoning clinic

Recruitment: children were recruited in July–October 1981 as they appeared for regular blood lead monitoring in a lead poisoning clinic

Follow‐up: 12 months from baseline

Intention‐to‐treat: unclear

Power calculation: performed to determine number of participants (unclear if required number recruited)

Participants

Eligibility criteria

1. BLL 30.0–49.0 µg/dL

2. Children from lead poisoning clinic in Baltimore

3. Aged 15–72 months

4. Lived at present address for ≥ 6 months

Participation rate: 78/78 (100%) children enrolled and randomised (22 intervention, 56 control)

Reasons for non‐participation: NA

Analysis: 49/78 (63%) analysed

Number of dropouts/withdrawals: 29

Reasons for dropout/withdrawal: children moved, spent considerable time with relatives in another household, not home for visits (no specific numbers per reason reported)

Intervention baseline characteristics (available for n = 14)

1. 36% boys, 64% girls

2. Mean age 45 months

3. Mean BLL 38.0 µg/dL

Control baseline characteristics (available for n = 35)

1. 51% boys, 49% girls

2. Mean age 43 months

3. Mean BLL 38.0 µg/dL

Interventions

Intervention

1. Dust control team to wet mop all rooms twice per month

2. Parental education to clean more frequently over 12‐month period

3. Paint stabilisation

Control

1. Routine advice dust control by mopping given at clinic plus paint stabilisation

Outcomes

1. BLL 12 months from baseline (venous blood sample)

2. Household floor dust (only reported in intervention group)

Notes

Funding: supported from HUD Grant

Conflicts of interest: none declared

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

High risk

Quote from report: [allocation method alternate based on] "even or odd clinic number."

Allocation concealment (selection bias)

High risk

Comment: not used.

Blinding of participants and personnel (performance bias): objective outcomes

Low risk

Quote from correspondence with author: "personnel was not blinded. Participants were not aware of the existence of another study group."

Comment: we rated this domain at low risk of bias because BLLs are unlikely to be influenced by participants' knowledge about treatment allocation.

Blinding of outcome assessment (detection bias)
All outcomes

Low risk

Comment: blinding of outcome assessors (laboratory).

Incomplete outcome data (attrition bias)
Blood lead level

Unclear risk

Comment: attrition rate was quite high (37%), but there was no difference in attrition rates between study arms (37% in both groups).

Selective reporting (reporting bias)

Unclear risk

Comment: insufficient information.

Other bias

Low risk

Comment: study appeared free from other sources of bias.

Study characteristics

Methods

Study design: cluster‐RCT, by neighbourhood

Study location/setting: Baltimore, USA. Urban neighbourhoods

Recruitment: children from 2 neighbourhoods were recruited beginning in 1988

Follow‐up: 1 year

Intention‐to‐treat analysis: no, due to exclusion of those not adhering to the study protocol (n = 226)

Power calculation: performed to determine number of participants (required number recruited)

Participants

Eligibility criteria for neighbourhoods

1. Sufficient children to test hypothesis

2. Areas of exposed soil around homes

3. Pre‐1950 urban housing away from major industries or highways

4. Comparable demographics

5. Moderate risk for lead exposure

Eligibility criteria for children

1. 6 months to 6 years of age

2. Living in the same house (in selected neighbourhood) for ≥ 3 months and family was not planning to move

Participation rate: NA as community recruitment; 408 children (212 intervention, 196 control) in 263 houses randomised

Reasons for non‐participation: NA

Analysis: 182/408 (121/263 households) (45%) analysed

Number of dropouts/withdrawals: 226

Reasons for dropouts/withdrawals: children did not complete the study protocol (no specific reasons reported)

Intervention baseline characteristics (n = 212): mean BLL 11.0 µg/dL

Control baseline characteristics (n = 196): mean BLL 10.9 µg/dL

No information on sex or age of included children

Interventions

Intervention

1. Soil abatement consisting of removing the top 6 inches (15 cm) of soil, replacing it with lead‐free soil, then sodding or seeding, all within 1 week of exterior paint stabilisation

2. External paint stabilisation as means of preventing soil recontamination

Control

1. External paint stabilisation as means of preventing soil recontamination

Outcomes

1. BLL at 2 years from baseline (venous blood sample)

2. Soil lead levels

Notes

Funding: US Environmental Protection Agency

Conflicts of interest: none declared

Other comments:

1. Baseline soil lead levels lower than hypothesised with 54% > 1000 parts per million

2. No internal household interventions

3. Adjacent properties not abated

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Low risk

Quote from correspondence with author: "coin toss."

Allocation concealment (selection bias)

Unclear risk

Comment: insufficient information.

Blinding of participants and personnel (performance bias): objective outcomes

Low risk

Comment: no information provided on blinding. We rated this domain at low risk of bias because BLLs and household dust lead measures are unlikely to be influenced by participants' knowledge about treatment allocation.

Blinding of outcome assessment (detection bias)
All outcomes

Low risk

Quote from correspondence with author: "Specimen collectors and laboratory personnel were blinded to group allocation and analyses were done by the State laboratory which had no interest in the outcome of the study."

Incomplete outcome data (attrition bias)
Blood lead level

High risk

Comment: attrition rate was very high (55%). Attrition was similar in both groups (intervention group 53%, control group 58%).

Selective reporting (reporting bias)

Unclear risk

Comment: insufficient information.

Other bias

Unclear risk

Comment: risk of bias due to cluster randomisation. Used neighbourhood clusters, and it was unclear how analysis was performed as data were NA – unclear if there is unit of analysis bias. We assessed risk of recruitment bias and bias risk due to baseline imbalance as low, since baseline characteristics were comparable and randomisation was achieved by coin toss.

Study characteristics

Methods

Study design: cluster‐RCT, by household (in blocks of 6 stratified by area and BLL)

Study location/setting: British Columbia, Canada. Higher lead risk area (active smelter)

Recruitment: blood screen in 1992

Follow‐up: 10 months from baseline

Intention‐to‐treat: available‐case analysis

Power calculation: performed to determine number of participants (required number recruited)

Participants

Eligibility criteria

1. Households in the study area with children < 72 months of age

2. No plans of moving

3. Living at the present address for > 1 month

Participation rate: 122/176 (69%) eligible households enrolled and randomised (122 children; 61 intervention, 61 control)

Reasons for non‐participation: 54 households were not interested in participating

Analysis: 111/122 (99%) analysed

Number of dropouts/withdrawals: 11

Reasons for dropouts/withdrawals: moved house or did not provide a final blood sample (no specific numbers per reasons reported)

Intervention baseline characteristics (available for n = 55)

1. Mean age 32.9 months

2. Mean BLL 11.9 µg/dL

Control baseline characteristics (available for n = 56)

1. Mean age 31.9 months

2. Mean BLL 11.3 µg/dL

3. No information on sex of children

Interventions

Intervention

1. HEPA vacuuming (7 times in a 10‐month period)

2. Routine advice regarding maintenance and general lead educational materials provided

Control

1. Routine advice regarding maintenance and general lead educational materials provided

Outcomes

1. BLL 10 months from baseline (venous blood sample)

2. Floor dust and lead levels

Notes

Funding: grants to the Trail Community Lead Task Force by: BC Ministry of Health, BC, Ministry of Environment, Lands and Parks, Cominco Limited and City of Trail

Conflicts of interest: none declared

Other comments: potential for unit of analysis error

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Low risk

Quote from correspondence with author: "drew concealed slips of paper numbered one to six without replacement" and assigned blocks and then "coin toss" determined that "odds would be treatment blocks."

Allocation concealment (selection bias)

Low risk

Quote from correspondence with author: "done in central office."

Blinding of participants and personnel (performance bias): objective outcomes

Low risk

Quote from correspondence with author: "participants and personnel were not blinded as to treatment allocation."

Comment: we rated this domain at low risk of bias because BLLs and household dust lead measures are unlikely to be influenced by participants' knowledge about treatment allocation.

Blinding of outcome assessment (detection bias)
All outcomes

Low risk

Quote from correspondence with author: "blood specimen collector and lab personnel did not know of group assignments", "lab personnel analysing the carpet dust samples were not aware of group assignment."

Incomplete outcome data (attrition bias)
Blood lead level

Low risk

Comment: low attrition rate (1%).

Incomplete outcome data (attrition bias)
Household dust lead level

Low risk

Comment: low attrition rate (1%).

Selective reporting (reporting bias)

Low risk

Comment: study protocol NA, but the published reports included all expected outcomes, including those that were pre‐specified and confirmed by the investigator.

Other bias

High risk

Comment: risk of bias for cluster‐randomised studies. Used clusters of 6 households but used individuals as unit for analysis and, therefore, introduced a unit of analysis error. We assessed risk of recruitment bias and bias risk due to baseline imbalance as low, since baseline characteristics were comparable and randomisation was achieved by coin toss.

Study characteristics

Methods

Study design: RCT

Study location/setting: Minneapolis, USA. Urban area

Recruitment: recruited by door knocking and community information

Follow‐up: 3 years from baseline

Intention‐to‐treat: unclear

Power calculation: performed to determine number of participants (unclear if required number recruited)

Participants

Eligibility criteria: pregnant women and mothers of young infants from the Phillips Neighbourhood (economically disadvantaged, ethnically diverse neighbourhood)

Participation rate: NA as community recruitment

Reasons for non‐participation: NA

Analysis: 607 children (299 intervention, 308 control) randomised, 378 (62%) analysed

Number of dropouts/withdrawals: 229

Reasons for dropouts/withdrawal: no reasons stated

No information of baseline data was available for all randomised children

1. According to the authors, there was no difference in baseline characteristics between treatment groups (no table presented, no detailed information on sex or age)

2. Mean BLL < 10.0 µg/dL

Interventions

Intervention

1. Intensive educational intervention: 20 bi‐weekly, culturally specific educational session by peer leaders provided individually and 3‐monthly boosters until child was 3 years of age

2. Routine state health brochures about lead, home assessment for lead contamination and feedback about home inspections

Control

1. Routine state health brochures about lead, home assessment for lead contamination and feedback about home inspections

Outcomes

BLL (capillary until 12 months, venous > 12 months) 3 years from baseline

Notes

Funding: supported by Grant MCJ 270801 from the Maternal and Child Health Bureau and Grant U67/CCU510771 from the CDC

Conflicts of interest: authors declared they had no competing financial interests

Other comments

1. Dichotomous data

2. All participants given financial incentive

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Low risk

Quote from correspondence with author: "Random number generator."

Allocation concealment (selection bias)

Low risk

Quote from correspondence with author: "Central office."

Blinding of participants and personnel (performance bias): objective outcomes

Low risk

Quote from correspondence with author: "no blinding of participants or personnel."

Comment: we rated this domain at low risk of bias because BLLs are unlikely to be influenced by participants' knowledge about treatment allocation.

Blinding of outcome assessment (detection bias)
All outcomes

Low risk

Quote from correspondence with author: outcome assessors "laboratory" blinded.

Incomplete outcome data (attrition bias)
Blood lead level

High risk

Comment: attrition was quite high (38%), no information on attrition rates in both groups were given, so we rated it at high risk of bias.

Selective reporting (reporting bias)

High risk

Comment: results on household dust lead outcomes were not reported although they were measured. No information from the author on these outcomes could be obtained.

Other bias

Low risk

Comment: study appeared free from other sources of bias.

Study characteristics

Methods

Study design: RCT

Study location/setting: Rochester, NY, USA. Community‐based trial in urban area

Recruitment: baseline data collected between August 1993 and November 1993; follow‐up samples collected between April 1994 and June 1994

Follow‐up: 7 months from baseline

Intention‐to‐treat available‐case analysis

Power calculation: not performed to determine number of participants

Participants

Eligibility criteria: families with children who participated in the Lead‐in Dust study (a cross‐sectional study to assess relationship of lead‐contaminated house dust and urban children's BLLs)

Participation rate: 104/205 (50%) enrolled (no significant difference in those refused) and randomised (57 intervention, 47 control)

Reasons for non‐participation: 101 not interested in participating

Analysis: BLL 96/104 (91%), non‐carpet floor dust lead level 70/104 (67%), carpet floor dust lead level 60/104 (57%)

Number of dropouts/withdrawals: 8 (BLL), 34 (non‐carpet floor lead dust level), 44 (carpet floor lead dust level)

Reasons for dropout/withdrawal: 2 refused second blood tests, 3 had moved outside of the area, 2 had no time and 1 was lost to follow‐up; for others of whom floor dust lead levels were NA and no reasons were specified

Intervention baseline characteristics (n = 57)

1. Mean age 19.8 months

2. Mean BLL 6.6 µg/dL

Control baseline characteristics (n = 47)

1. Mean age 20.4 months

2. Mean BLL 6.8 µg/dL

No information on sex of participants

Interventions

Intervention

1. Trained interviewer emphasised the importance of dust control for reducing children's exposure to lead, provided them with cleaning supplies, gave a demonstration of how to clean, and instructed families how and when to clean

2. Families were given a colouring book that described lead poisoning and its prevention

3. Families were given a brochure on lead poisoning and its prevention

Control

1. Families with children who participated in the Lead‐in Dust study (a cross‐sectional study to assess relationship of lead‐contaminated house dust and urban children's BLLs)

Outcomes

1. BLL at 7 months from baseline (venous blood sample)

2. Household floor and window dust lead

Notes

Funding: grant NYLPR002‐94 from the US HUD, the National Center for Lead‐Safe Housing, and Institutional National Research Service Award 2T‐32 PE‐12002 from the Bureau of Health Professions, Health Resources and Services Administration, US Public Health Service, Department of Health and Human Services

Conflicts of interest: none declared

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Low risk

Quote from correspondence with author: "computer random number generator."

Allocation concealment (selection bias)

Low risk

Quote from correspondence with author: "sealed opaque envelopes."

Blinding of participants and personnel (performance bias): objective outcomes

Low risk

Quote from correspondence with author: "personnel and participants were blinded."

Blinding of outcome assessment (detection bias)
All outcomes

Low risk

Quote from correspondence with author: "yes, blood lead specimen collectors and analysers were blinded to group allocation."

Incomplete outcome data (attrition bias)
Blood lead level

Low risk

Comment: attrition rate for BLL was low (9%).

Incomplete outcome data (attrition bias)
Household dust lead level

Low risk

Comment: attrition rate was acceptable (33%) and similar between groups (intervention group 32%, control group 36%).

Selective reporting (reporting bias)

Low risk

Comment: study protocol is available and all the study's prespecified outcomes were reported in the prespecified way.

Other bias

Low risk

Comment: study appeared free from other sources of bias.

Study characteristics

Methods

Study design: RCT (also non‐study control to rule out Hawthorne effect)

Study location/setting: Rochester, NY, USA. Urban area

Recruitment: identified by birth data from hospitals, inner‐city clinics and the Department of Social Services and Health – families were called to determine eligibility via interviews

Follow‐up: 42 months from baseline

Intention‐to‐treat: available‐case analysis

Power calculation: performed to determine number of participants (required number recruited)

Participants

Eligibility criteria

1. Living in Rochester, NY

2. No plans to relocate in next 3 months

3. Children aged > 5 months but < 7 months of age at baseline visit

Participation rate: 275/429 (64%) enrolled and randomised (140 intervention, 135 control)

Reasons for non‐participation: not interested in participating

Analysis: 245/275 (89%) analysed at 24 months and 189/275 (69%) analysed at 48 months

Number of dropouts/withdrawals: 30 at 24 months, 86 at 48 months

Reasons for dropout/withdrawal: lost to follow‐up

Intervention baseline data (n = 140)

1. Mean age 6.68 months

2. Mean BLLs 2.8 µg/dL

Control baseline data (n = 135)

1. Mean age 6.65 months

2. Mean BLL 2.9 µg/dL

No information on sex of participants

Interventions

Intervention

1. Up to 8 visits by dust control advisors, cleaning equipment and supplies in 24‐month period. Dust control advisors were trained to use educational model developed specifically for home visitation

2. Baseline 4 home visits by trained interviewer to collect data

Control

1. Baseline 4 home visits by trained interviewer to collect data

2. Families in the control group did not receive any lead exposure prevention education or intervention

Outcomes

1. BLL measured at 6 months (baseline), 12, 18, 24, 36 and 48 months (venous blood sample)

2. Household floor and window dust lead

Notes

Funding: CDC Grant (U67/CCU210773) and an Institutional National Research Service Award (#2T‐32 PE‐12002) from the Bureau of Health Professions, Human Resources and Services Administration, Public Health Service, Department of Health and Human Services

Conflicts of interest: none declared

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Low risk

Quote from correspondence with author: "random number generator."

Allocation concealment (selection bias)

Low risk

Quote from correspondence with author: "sealed opaque envelopes."

Blinding of participants and personnel (performance bias): objective outcomes

Low risk

Quote from correspondence with author: "personnel and participants were blinded."

Blinding of outcome assessment (detection bias)
All outcomes

Low risk

Quote from correspondence with author: "yes, blood lead specimen collectors and analysers were blinded to group allocation", "environmental technicians and interviewers blind to group assignment."

Incomplete outcome data (attrition bias)
Blood lead level

Low risk

Comment: attrition rate at 12 months was 11%, at 24 months it was 31% – both acceptable. Attrition rate between groups was similar.

Incomplete outcome data (attrition bias)
Household dust lead level

Low risk

Comment: attrition rate at 12 months was 11%, at 24 months it was 31% – both acceptable. Attrition rate between groups was similar.

Selective reporting (reporting bias)

Low risk

Comment: study protocol is available and all the study's prespecified outcomes were reported in the prespecified way.

Other bias

Low risk

Comment: study appeared free of other sources of bias.

Study characteristics

Methods

Study design: RCT

Study location/setting: medium‐sized, mid‐western city in the USA

Recruitment: low‐income families recruited from Women, Infants, and Children (n = 29), Head Start (n = 49) and the local Health Department (n = 6)

Follow‐up: 6 months from baseline

Intention‐to‐treat: no

Power calculation: not reported

Participants

Eligibility criteria for children

1. Raised up in low‐income families

2. Aged < 6 years

3. BLL 3–9.9 µg/dL

Participation rate: 332 children assessed for eligibility, 84 randomised

Reason for non‐participation: 26 refused to participate, 2 consented to participate but never completed initial interview, 203 disconnected/never contacted

Analysis: 77/84 (92.7% analysed)

Number of dropouts/withdrawals: 7/84 (7.3% dropout)

Reasons for dropout/withdrawal: 5 lost contact/not compliant, 1 did not want to participate, 1 unable to contact after relocating

Baseline characteristics of interventions and active control group (available for n = 84)

1. Mean age 47.3 months (SD 18.12)

2. BLL 5.28 µg/dL (SD 1.85)

No information on sex of children

Interventions

Intervention 1

1. Participants received a cleaning kit, including a Riccar Radiance HEPA vacuum cleaner ("cleaning kit", n = 21)

Intervention 2

1. Participants received verbal and written instructions on how to properly clean to reduce lead or a professional home inspection for lead and consultation outlining specific risks present and steps to alleviate risks ("risk assessment", n = 20)

Intervention 3

1. Participants received both a cleaning kit and verbal and written instructions or a professional home inspection for lead and consultation ("cleaning kit and risk assessment", n = 21)

Control

1. Education on lead poisoning through brochures (n = 22)

Passive control group (n = 29)

1. Data for this group were retrospectively retrieved from charts and not gathered prospectively within the RCT; therefore, data on this group was not used for this review.

Outcomes

1. BLLs at baseline and 6 months

Notes

Funding: provided by grants from the University of Notre Dame's Rodney Ganey Community‐Based Research mini‐grant, the New York Community Trust's Fahs‐Beck Fund for Research and Experimentation, South Bend Housing and Urban Development lead hazard control grant, and private donations from Greentree Environmental, LLC, and Riccar, Inc.

Conflicts of interest: not reported

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Low risk

Quote from report: "Sealed envelopes contained the name of each of the four intervention groups, which were shuffled to ensure random order." (p. 381).

Allocation concealment (selection bias)

Low risk

Comment: interviewers selected 1 of the sealed envelopes prior to the initial interview but opened it afterwards in front of the participants.

Blinding of participants and personnel (performance bias): objective outcomes

Low risk

Comment: not blinded according to personal communication with the author. We rated this domain at low risk of bias because BLLs are unlikely to be influenced by participants' knowledge about treatment.

Blinding of outcome assessment (detection bias)
All outcomes

Unclear risk

Comment: not stated.

Incomplete outcome data (attrition bias)
Blood lead level

Low risk

Comment: attrition rate 14.7% and low differential attrition.

Selective reporting (reporting bias)

Unclear risk

Comment: study protocol NA.

Other bias

Low risk

Comment: study appeared free from other sources of bias.

Study characteristics

Methods

Study design: RCT

Study location/setting: Jersey City, NY, USA

Recruitment: families that responded to posters and door hangers or were referred to the study by the municipal lead poisoning prevention programme, local healthcare providers, or recommendation.

Follow‐up: 12 months from baseline

Intention‐to‐treat: unclear

Power calculation: performed to determine number of participants (required number not recruited)

Participants

Eligibility criteria

1. Children aged 6 months to 3 years

2. Responsible adult had to speak English or Spanish

3. Presence of lead paint in the home

4. Home had to be in a state that could be cleaned effectively (not structural disrepaired)

5. No evidence of illicit drug use, firearms or other major staff safety concerns

6. Index child was not in regular day care

Participation rate: 113/147(77%) enrolled and randomised (56 intervention, 57 control)

Reasons for non‐participation: 7 could not be recontacted or refused to allow a baseline blood lead sample to be drawn, 27 were not interested in participating.

Analysis: 99/113 (87%) analysed for BLLs, 95/113 (84%) analysed for floor wipes, 76/113 (67%) analysed for sill wipes, 49/113 (43%) analysed for vacuum cleaning

Number of dropouts/withdrawals: 14 BLL, 18–64 household dust lead levels

Reasons for dropout/withdrawal: because of frequent moves and changing circumstances of the enrolled families, it was not possible to draw final blood samples from 14 children. No explanation for missing data on household dust lead levels stated

Intervention baseline characteristics (available for n = 46)

1. Mean age 1.7 years

2. Mean BLL 12.4 µg/dL

Control baseline characteristics (available for n = 53)

1. Mean age 1.6 years

2. Mean BLL 11.6 µg/dL

No information on sex of children

Interventions

Intervention

1. Bi‐weekly assistance with household cleaning (HEPA vacuum cleaner and wet mopping) by community staff members for 1 year. Visits usually lasted 2 hours.

2. Offer to attend 4 or 5 educational sessions a year about lead prevention

Control

1. Accident prevention group given household safety items, but no assistance with household cleaning and no special education on lead prevention during visits

2. Offer to attend 4 or 5 educational sessions a year about lead prevention

Outcomes

1. BLL 12 months from baseline

2. Household dust lead levels

3. Maternal lead knowledge

Notes

Funding: work was supported by Co‐operative Agreement CR820235 from the Environmental Protection Agency, by an Interagency Agreement from the National Institute for Child Health and Human Development, National Institutes of Health, to US Environmental Protection Agency, by Grant 18152 from the Robert Wood Johnson Foundation, and by Grant ES‐05022 from the National Institute of Environmental Health Sciences.

Conflicts of interest: none declared

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Low risk

Quote from correspondence with author: "permutated blocks of varying length."

Allocation concealment (selection bias)

Low risk

Quote from correspondence with author: "sealed envelopes."

Blinding of participants and personnel (performance bias): objective outcomes

Low risk

Quote from correspondence with author: "it was not possible to blind participants or field personnel to the assignments since one group had cleaning teams come to their homes and the other group did not."

Comment: we rated this domain at low risk of bias because BLLs and household dust lead measures are unlikely to be influenced by participants' knowledge about treatment allocation.

Blinding of outcome assessment (detection bias)
All outcomes

Low risk

Comment: outcome assessors were blinded.

Incomplete outcome data (attrition bias)
Blood lead level

Low risk

Comment: low attrition rate (12%).

Incomplete outcome data (attrition bias)
Household dust lead level

Unclear risk

Comment: for the outcome household dust, attrition rate was higher and reasons for missing data were NA.

Selective reporting (reporting bias)

Unclear risk

Comment: insufficient information.

Other bias

Low risk

Comment: study appeared free from other sources of bias.

Study characteristics

Methods

Study design: RCT

Study location/setting: not reported (author's affiliation is Shanghai)

Recruitment: children's blood was tested for lead and children with confirmed BLL ≥ 10 μg/dL were included. Recruitment NA.

Follow‐up: 3 months

Intention‐to‐treat: not necessary, all 200 participants randomised remained in study

Power calculation: not reported

Participants

Eligibility criteria: children with confirmed BLL ≥ 10 μg/dL

Participation rate: 100%

Reason for non‐participation: not reported

Analysis: 200/200

Number of dropouts/withdrawals: 0/200

Reasons for dropout/withdrawal: no dropouts

Intervention baseline characteristics (available for n = 107)

1. 59.8% boys, 40.2% girls

2. Mean age 48.9 months (SD 15.7)

3. BLL 15.9 µg/dL (SD 9.7)

Control baseline characteristics (available for n = 93)

1. 58.1% boys, 41.9% girls

2. Mean age 48.6 months (SD 18.1)

3. BLL 16.6 µg/dL (SD 7.8)

Interventions

Intervention

1. Parental education undertaken by means of a television programme, a set of slides and a brochure, and focused on the questions regarding harmful effects of lead poisoning, the sources of environmental lead and prevention of domestic lead exposure.

Control

1. No education and no contact until end of study

Outcomes

1. BLL at baseline and after 3 months

Notes

Funding: Johnson & Johnson Research Foundation, USA

Conflicts of interest: not reported

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Unclear risk

Quote from the translated report: "All children were randomly divided into experimental group and control group according to their natural classes or groups in the kindergarten" (p. 893).

Allocation concealment (selection bias)

Unclear risk

Comment: not stated.

Blinding of participants and personnel (performance bias): objective outcomes

Low risk

Comment: not stated. We rated this domain at low risk of bias for objective outcomes because BLLs are unlikely to be influenced by participants' knowledge about treatment.

Blinding of outcome assessment (detection bias)
All outcomes

Low risk

Quote from the translated report: "…the laboratory testers did not know the grouping situation" (p. 893).

Incomplete outcome data (attrition bias)
Blood lead level

Low risk

Comment: no dropouts.

Selective reporting (reporting bias)

Unclear risk

Comment: study protocol NA.

Other bias

Low risk

Comment: study appeared free from other sources of bias.

Study characteristics

Methods

Study design: RCT

Study location: Missouri, USA. Former lead mining areas with large amounts of lead mining waste

Recruitment: screenings in health department lead clinics; women, infants and children's clinics; daycare centres; door‐to‐door screening; and health fairs

Follow‐up: 9 months

Intention‐to‐treat: only cases with all quarterly measurements were included in analysis

Power calculation: performed to determine number of participants (unclear if required number recruited)

Participants

Eligibility criteria

1. Children 6–72 months of age

2. BLL 10.0–20.0 µg/dL

3. For 1 of the 2 counties, the households were required to be below the medium income level for the area

Participation rate: 101/134 (75%) randomised (34 intervention 1, 35 intervention 2, 32 control)

Reasons for non‐participation: no reasons for non‐participation provided

Analysis: BLL 39/101 (39%), household lead dust 34/101 (34%) analysed

Number of dropouts/withdrawals: 62 BLL, 67 household lead dust level

Reasons for dropout/withdrawal: no reasons for dropouts provided

Intervention 1 baseline characteristics (n = 34)

1. 59% boys, 41% girls

2. Mean age 2.8 years

3. Mean BLL 12.8 µg/dL

Intervention 2 baseline characteristics (n = 35)

1. 49% boys, 51% girls

2. Mean age 3.6 years

3. Mean BLL 12.7 µg/dL

Control baseline characteristics (n = 32)

1. 47% boys, 53% girls

2. Mean age 2.8 years

3. Mean BLL 12.7 µg/dL

Interventions

Intervention 1

1. A standard health education session on lead exposure reduction activities, given by a nurse educator, a letter reporting the results of the environmental lead assessment of the home, generic educational information in the form of pamphlets produced by state and federal agencies

2. 3 × quarterly educational home visit by nurse (providing education on hygiene, nutrition, blood‐lead screening, house cleaning and providing cleaning supplies) and 6 personalised newsletters over 9‐month period

Intervention 2

1. As intervention 1, plus 3 × quarterly professional cleans with wet mopping, HEPA and carpet shampooing

Control

1. Standard health education session on lead exposure reduction activities, given by a nurse educator, a letter reporting the results of the environmental lead assessment of the home, generic educational information in the form of pamphlets produced by state and federal agencies

Outcomes

1. BLL at 3, 6 and 9 months from baseline (not stated if venous or capillary sample)

2. Household lead dust levels until 9 months from baseline (only presented graphically

Notes

Funding: not stated

Conflict of interests: none declared

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Unclear risk

Comment: method of randomisation unknown.

Allocation concealment (selection bias)

Unclear risk

Comment: unknown.

Blinding of participants and personnel (performance bias): objective outcomes

Low risk

Quote from correspondence with author: "no blinding of participants or researchers occurred."

Comment: we rated this domain at low risk of bias because BLLs and household dust lead measures are unlikely to be influenced by participants' knowledge about treatment allocation.

Blinding of outcome assessment (detection bias)
All outcomes

Low risk

Comment: outcome analysers (laboratory) blinded.

Incomplete outcome data (attrition bias)
Blood lead level

High risk

Comment: high attrition rate (61%). Reasons for missing data NA.

Incomplete outcome data (attrition bias)
Household dust lead level

High risk

Comment: high attrition rate (66%). Reasons for missing data NA.

Selective reporting (reporting bias)

Unclear risk

Comment: insufficient information.

Other bias

Low risk

Comment: study appeared free from other sources of bias.

Study characteristics

Methods

Study design: RCT

Study location/setting: Florida, USA

Recruitment: carers selected from clients enrolled in Broward County MediPass (Medicaid) who selected Children's Diagnostic and Treatment Center as their healthcare provider

Follow‐up: 4 months from baseline

Intention‐to‐treat: unclear

Power calculation: not performed to determine number of participants

Participants

Eligibility criteria

1. Aged 1–3 years

2. MediPass (Medicaid) as their insurance

Participation rate: 63/63 (100%) children randomised (32 intervention, 31 control)

Reasons for non‐participation: NA

Analysis: 50/63 (79%) analysed

Number of dropouts/withdrawals: 13

Reasons for dropout/withdrawal: reasons not specified

Intervention baseline characteristics (available for n = 28)

1. Mean age 23.5 months

2. Mean BLL 4.5 µg/dL

Control baseline characteristics (available for n = 22)

1. Mean age 21.5 months

2. Mean BLL 2.6 µg/dL

Sex of children not reported

Interventions

Intervention

1. Education session at clinic consisting of print‐based module written by the researcher and used as the basis of parental lead education, a video used to show methods parents could use in the home to prevent lead poisoning, and brochure highlighting the risks of childhood lead exposure including factors that affect the home environment, behaviours that mitigate risk, and the need for proper nutrition at first clinic

Control

1. Education session as described above at second clinic (wait‐list control)

Outcomes

1. BLL at 3–4 months from baseline (venous blood sample)

2. Parental knowledge – Chicago Lead Knowledge Test

Notes

Funding: not stated

Conflicts of Interest: none declared

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Low risk

Quote from correspondence with author: "random list of numbers."

Allocation concealment (selection bias)

Low risk

Quote from correspondence with author: "assigned by central office."

Blinding of participants and personnel (performance bias): objective outcomes

Low risk

Quote from correspondence with author: "Personnel was not blinded. Participants were unaware of the existence of the comparison group."

Comment: we rated this domain at low risk of bias because BLLs are unlikely to be influenced by participants' knowledge about treatment allocation.

Blinding of outcome assessment (detection bias)
All outcomes

Low risk

Comment: blinding of outcome assessors (laboratory).

Incomplete outcome data (attrition bias)
Blood lead level

High risk

Comment: attrition rate (21%) was acceptable. However, attrition rate was much higher in control group (30%) than in intervention group (12%). Reasons for dropouts not specified.

Selective reporting (reporting bias)

Low risk

Comment: study protocol NA but the published reports included all expected outcomes, including those that were prespecified and confirmed by the investigator.

Other bias

Low risk

Comment: study appeared free from other sources of bias.

Study characteristics

Methods

Study design: RCT

Study location/setting: Boston, USA. Urban neighbourhoods with a high incidence of childhood lead poisoning and high soil lead levels

Recruitment: screening efforts between January and June 1989

Follow‐up: 11 months from baseline

Intention‐to‐treat: available‐case analysis

Power calculation: performed to determine number of participants (required number recruited)

Participants

Eligibility criteria

1. Aged ≤ 4 years

2. Finger‐stick BLL of 9.94–19.88 µg/dL

3. Chipping or peeling paint did not exceed 30% of the total surface area on the exterior walls of the child's home or exceed 40% on the walls of abutting premises

4. Premise had a yard of ≥ 0.9 m² composed of dirt or grass, or both, that was accessible to the child and the mean or median surface soil lead level among samples taken near the house was ≥ 1500 parts per million

5. Child resided in a dwelling with ≤ 8 residential units and was mobile

6. Child had never been lead poisoned

7. Family resided on the premises for ≥ 3 months and had no plans to move within the 3 months after enrolment

Participation rate: 152/236 (64%) children randomised (54 intervention, 51 control group 1, 47 control group 2)

Reasons for non‐participation: children who had venous BLLs > 1.16 µmol/L were excluded because they met the former definition of lead poisoning and were likely to undergo medical and environmental interventions that could obscure changes associated with the study interventions (not specified whether all 84 children had high BLLs or if other reasons account for non‐participation)

Analysis: 149/152 (98%) analysed

Number of dropouts/withdrawals: 3

Reasons for dropouts/withdrawals: no specific reasons for dropouts stated

Intervention baseline characteristics (available for n = 52)

1. 60% boys, 40% girls

2. Mean age 30.5 months

3. Mean BLL 0.6 µg/dL

Control group 1 baseline characteristics (n = 51)

1. 49% boys, 51% girls

2. Mean age 31.4 months

3. Mean BLL 0.6 µg/dL

Control group 2 baseline characteristics (n = 47)

1. 51% boys, 49% girls

2. Mean age 33.1 months

3. Mean BLL 0.6 µg/dL

Interventions

Phase I only
Intervention

1. Soil abatement from yard (15 cm layer of topsoil was removed and replaced with 20 cm of clean soil)

2. Interior dust abatement (HEPA filter vacuum cleaning and wiping surfaces with a wet cloth or an oil‐treated rag for furniture, floors, walls, woodwork, windows, furniture surfaces were cleaned)

3. Loose interior paint removal (vacuuming the loose paint areas with HEPA, washing loose paint areas, painting window wells with primer)

Control 1

1. Interior dust abatement (HEPA filter vacuum cleaning and wiping surfaces with a wet cloth or an oil‐treated rag for furniture, floors, walls, woodwork, windows, furniture surfaces were cleaned)

2. Loose interior paint removal (vacuuming the loose paint areas with HEPA, washing loose paint areas, painting window wells with primer)

Control 2

1. Loose interior paint removal (vacuuming the loose paint areas with HEPA, washing loose paint areas, painting window wells with primer)

Outcomes

1. BLLs 11 months from baseline (venous blood sample)

2. Household dust levels (only reported as percentage of households that remained at a lower dust level than at baseline after 4–5 weeks and after 33 weeks after the intervention)

Notes

Funding: grant X001822‐01‐06 from Environmental Protection Agency, Washington, DC

Conflicts of interest: none declared

Other comments: phase I and phase II of Boston Lead‐In‐Soil trial performed but phase II excluded as no controls

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Low risk

Quote from correspondence with author: "computer‐based random number generator."

Allocation concealment (selection bias)

High risk

Comment: allocation performed by 1 staff member but not actively concealed from other investigators enrolling participants.

Blinding of participants and personnel (performance bias): objective outcomes

Low risk

Quote from correspondence with author: "impossible for participants or personnel to be blinded to treatment allocation."

Comment: we rated this domain at low risk of bias because BLLs and household dust lead measures are unlikely to be influenced by participants' knowledge about treatment allocation.

Blinding of outcome assessment (detection bias)
All outcomes

Low risk

Comment: outcome assessors (laboratory analysers) were blinded.

Incomplete outcome data (attrition bias)
Blood lead level

Low risk

Comment: attrition rate was low (2%).

Incomplete outcome data (attrition bias)
Household dust lead level

Low risk

Comment: attrition rate was low (2%).

Selective reporting (reporting bias)

Unclear risk

Comment: insufficient information.

Other bias

Low risk

Comment: study appeared free from other sources of bias.