Effectiveness of educational interventions for the prevention of lead poisoning in children: a systematic review

Literature search

This review followed the Preferred Reporting Items for Systematic review and Meta-Analysis Protocols (PRISMA-P) guidelines [35], a set of evidence-based standards for reporting systematic reviews. A search strategy was created in partnership with a medical librarian, incorporating database specific subject headings and keywords related to the topics of lead poisoning and education. Search terms related to lead included “lead”, “blood lead level”, and the elemental form of lead, “Pb”. Search terms related to education included: “health education”, “promotion”, and “health fair”. Searches were conducted in June of 2021 on four databases: Ovid MEDLINE, Scopus, Web of Science Core Collection, and CINAHL (EBSCO). Search results were updated November 2023 to incorporate results from the most recent two years, at that time one additional article met criteria for inclusion. See Appendix A for full search strategy.

Article selection process

Duplicates were removed and search results were exported to a web-based application for systematic reviews, Rayyan [36]. Two independent reviewers (JSB, JCB) screened the results, iteratively applying a systematic PICOS criterion [37] for inclusion/exclusion:

Eligibility criteria

Studies were included if they were an experimental or quasi-experimental design. Studies were included if they described an educational intervention, even if education was just one component of a multi-arm intervention. Studies were excluded if they were not original research, such as systematic reviews. Studies were excluded if they did not describe an educational intervention aimed at reducing or preventing lead poisoning in children. Finally, studies were excluded if the outcome was not BLL in children.

Demographics

Most of the articles (12/17) studied children residing in the United States. Seven of the studies were based in the east/northeast areas of the country: Brown et al. looked at a sample of 175 children <28 months of age in Rhode Island [38]. Two studies (and a follow-up) were based in Philadephphia, PA: Campbell’s original and follow-up study looked at a sample of 314 newborns, and the Rothman et al. study of an underserved African-American population [39], [40], [41]. Lanphear et al. analyzed a group of 104 families with children aged 12–31 months in Rochester, NY [42]. Rhoads et al. studied a sample of 113 children from Jersey City, NJ [43]. The study by Schucker et al. of children <4 years old took place in Baltimore, MD [44].

Five studies were conducted in the midwest: The Jordan et al. study looked at a sample of mothers with children aged 0–36 months in Minneapolis, MN [45]. Kassa et al. looked at a sample of children in Toledo, OH [46]. The Nicholson study looked at a group of 84 children from low income families from a “medium-sized midwestern city” [47]. Two studies took place in Milwaukee, WI: The Schlenker et al. study of 1850 children aged 6 months–6 years, and the Schultz et al. study of 413 children [48], 49]. The Kimbrough et al. study took place in Illinois in the vicinity of a former lead smelting site [50].

Only one study was conducted in the southern region of the United States: the study by Sterling et al. of 134 children aged 6–72 months took place in two counties in Missouri, both former lead mining areas [51]. No studies or samples from the western region United States were represented.

Three studies were conducted outside of the United States. The study by Burke et al. surveyed 7,885 children (with estimated response rate between 41 and 75 %) living in a former lead mining/high exposure area took place in Broken Hill, Australia [52]. The Hilts et al. study took place in Trail, Canada-also a site of an active lead/zinc smelter and high risk area [53]. The study by Morales et al. took place in the Callao region of Peru, and area also prone to contamination by local smelters [54].

In four of the studies, children were at environmental risk of lead exposure due to being located at sites or former sites of lead mining/smelting or other industries [50], [52], [53], [54].

Study design

Eight of the studies (including a follow-up study) compared their intervention group to a control group using a randomized controlled design. Brown et al. compared their study group with a group who received customary care (provided standard health education about lead poisoning with less visits and customization than intervention group) [38]. Sterling et al. compared their intervention group to a group receiving “standard” care which was comprised of one 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; and generic educational information in the form of pamphlets produced by state and federal agencies [51]. In their study and follow up analysis, Campbell et al. compared their intervention group against a control group that also received standard lead education in their original and follow-up study [39], 40]. Jordan et al., Lanphear et al., and Nicholson compared their intervention group to a control group who received a brochure on basic lead prevention [42], 45], 47]. Rhoads et al. compared their intervention group to control group that received an ‘accident prevention’ intervention, which notably did include some lead education [43].

Four used prospective or retrospective cohort-style designs to test their intervention. Rothman et al. compared census tracts in which people received the intervention to similar census tracts that did not receive the intervention [41]. Schlenker et al. compared their intervention group to a comparison group of BLL data for the city by zip code [49]. Schucker et al. compared tracts in Baltimore that received the intervention with control tracts [44]. The study by Schultz et al. compared their intervention group to a non-intervention control group retrospectively. The reference group presumably was comprised of children who like the experiemental group had BLLs between 20 and 24 μg/dL in the Milwaukee Health department records, but did not receive the educational intervention [48].

Five studies used a single cohort with a pre/post test to study their intervention: Burke et al., Hilts et al., Kassa et al., and Kimbrough et al., and Morales et al. [46], 50], [52], [53], [54].

Interventions

As it was the focus of this review, all studies included an educational component, with or without additional intervention components. Four studies used a solely educational intervention. Jordan et al. provided a brochure and offered bi-weekly educational sessions by same-ethnicity peer educators over a year, with quarterly booster sessions thereafter [45]. Their bi-weekly curriculum included education on lead sources, strategies to reduce exposure, and nutrition. The intervention in the study by Schucker et al. provided in-home education by same language community health workers, during which surfaces were inspected with an adult in charge of the children’s care present, information on the dangers of lead poisoning was presented to the same adult, a leaflet was left with the family. An additional five visits at six months intervals were held to remind parents of lead poisoning dangers, details not provided on if the educator provided personalized education [44]. The intervention described in the Schultz et al. study included an hour-long educational visit from a paraprofessional emphasizing lead exposure, nutrition, and cleaning practices and printed materials were left with the family [48]. The Kimbrough et al. study provided in-home counseling for families in 30–45 min sessions on sources of lead, hygiene, housekeeping, and nutrition, and written educational materials were provided [50]. Kimbrough et al. also provided some environmental testing at baseline, though it is unclear if families received these results during the intervention period.

The remaining studies used a variety of additional interventions, such as: dust/soil testing [38], 51], exposure assessments/home evaluations [[38], [39], [40, 46], 47], 49], case finding/monitoring [52], remediation of public land [52], lead or lead dust remediation [39], 40], 46], 53], case management [53], chelation therapy [46], distribution of nutritional supplementation [54], distribution of cleaning supplies [[39], [40], [41], [42, 47], 51], 53], cleaning demonstrations [42], 49], 51], assistance with household cleaning or cleaning services [43], 53], provision of testing swabs for community members to test their residences [41], and assessment of parent child-interaction assessment [38] or occupational or recreational exposures [38], [39], [40]. In some cases where testing of soil, surfaces, or home evaluations were performed, the line between educational/non-educational interventions were blurry. For this review, inspections that appeared more formal, such as if the family was provided specific testing results so they knew with some certainty if there were hazards in their home and what they were specifically, were considered to be beyond education. Inspections that seemed less formal, or did not provided specific results but seemed to be used to personalize education, were considered educational.

Findings

Overall, findings on the success of educational interventions to reduce BLL in children were mixed. Four of five studies that used a pre/post design reported improvements in BLL during the treatment/intervention period. Burke et al. reported a decrease in age-sex standardized mean BLL from 16.7  to 7.0 μg/dL during the treatment period from 1999 to 2002 [52]. Hilts et al. reported changes in BLL of +0.5 to −4.0 μg/dL of those who received specific interventions, with statistically significant BLL decreases in three years out of five-although it is unclear how much decline can be attributed to the intervention vs. the overall decrease seen in general population [53]. Kimbrough et al. reported that for those who had a BLL ≥10 μg/dL at the start of the intervention, the arithmetic mean BLL four months later decreased from 15 to 7.8 μg/dL (although an additional sample a year later was taken of a subset that showed the mean increased again to 9 μg/dL) [50]. The study by Kassa et al. used a pre-post design and found that among the children in their study who received chelation therapy, the post-chelation mean BLL (expectedly) was lower although it remained elevated. Among nonchelated study participants who received the educational visit, the mean BLL was actually higher than their original screen [46]. Morales et al. reported a decline in mean BLL from 13.96  to 8.96 μg/dL. While it is promising that an initial decrease in BLL was demonstrated in these studies, none of these studies utilized control groups, and so it is difficult to discern if the decrease in BLL is due to the intervention, the general decline in BLL that has happened over time, or some other outside influence.

Jordan et al., Nicholson, Rhoads et al., Rothman et al., Schlenker et al., and Schultz et al. reported a decrease in BLL in their treatment group when compared to a control group. Jordan et al. reported that intervention participants were more likely to maintain BLL<10 μg/dL than the control group (81 vs. 73 % p=0.08), with multivariate analysis indicating that the intervention reduced the risk of BLL>10 μg/dL by 34 % [45]. Nicholson did not find significant differences between the different interventions in their study, but when they compared their intervention groups with a post-hoc passive control, there was a statistically significant improvement in BLL in all intervention groups [47]. Rhoads et al. reported that BLL fell 17 % in the intervention group but did not change in the control group that was given education unrelated to lead [43]. Schlenker et al. reported that those in their intervention group demonstrated a decreased rate in elevated BLL 1.6 times that of the city-wide average (p=0.016) [49]. Schultz reported that their intervention group had a decline in BLL 3.1 (15 %) higher than the reference group (p=0.001) [48]. The study by Rothman et al. reported that one-year post intervention there was a 9 % decline in BLL≥10 μg/dL in the intervention group, and only a 3 % decline in the control group [41].

Not all studies found that their interventions were effective. The studies by Brown et al., Campbell et al., the Campbell et al. two-year follow up study, Lanphear et al., Schucker et al. and Sterling et al. reported that there was no significant difference in BLL reduction between control/intervention groups [38], [39], [40, 42], 44], 51].

Some studies reported additional outcomes that may be mediating factors to the final outcome of decreased BLL, such as increased knowledge or improved environmental screening results (such as reduced lead dust). These mediating outcomes are not discussed further as they are not the focus of the review, but those findings are noted in Table 1.

Conclusions

Lead poisoning is a completely preventable condition with irreversible impacts on children who experience it. While this review demonstrated that there is a potential benefit to educational and combined interventions in the prevention of lead poisoning, mixed outcomes found in this review indicate that these interventions are not consistently effective at decreasing BLL in children. Interventions that had some measure of success in lowering BLL when compared against control groups included some notable elements: repeated or frequent educational visits, consideration of language, culture, ethnicity, and literacy, use of a community-based approach, home-based or individualized education, and provision of supplies or actual assistance with cleaning. While these are notable elements of the successful studies, one must also consider that some of these elements were also incorporated in the less-successful studies, making identification of the key drivers of success in these interventions difficult. Table 2 provides a brief summary of key components of interventions with results.

Comparison of effectiveness of educational interventions among studies was difficult due to the inability to assess the quality of the educational intervention (aside from studies who described potentially quality-impacting factors such as length of session or consideration of language, culture, and literacy). For interventions that were not exclusively educational, it is difficult to differentiate the impact of the educational component vs. the other components. The study populations were also varied, with differing age groups and differing baseline BLLs and among the children and differing study locations perhaps impacting outcomes.

Among the studies included, even the process of determining if a study was “successful” was somewhat subjective. While four of five studies that uses a pre-post design successfully demonstrated a decrease in BLL, it is hard to determine how much of the success can be attributed to the study intervention or simply the declining trend in BLL over time [55], or some other factor, such as local policy that reduced lead emissions in the local area [54]. Perhaps the same intervention compared to a control group would not have yielded a difference in rate or level of BLL decline. Additionally, among the randomized controlled studies, for ethical reasons, the control group often included some form of lead education, although to a lesser extent than the experimental arms. While many did not find significant differences between study groups, perhaps they would have found a significant different when compared to a group that received no intervention at all, as was the case in Nicholson’s 2008 study [47].

Limitations of this study include the potential to inadvertently exclude relevant articles during the search or screening process. Authors writing on this topic who did not have positive results when studying an intervention may have had more difficulty publishing, and therefore skewed the overall results of our review. Additionally, the results of this review are impacted by the overall burden of limitations among the articles, including lack of control group, variations in interventions and intervention quality, lack of statistical analysis, and the inability to precisely define the impact of the intervention vs. other outside influences.

Implications

This review has found that educational interventions do not consistently reduce BLL in children. Even the most effective interventions evaluated in this review did not decrease childhood lead exposure to an acceptable level: none. Although some studies did show some level of success in decreasing BLL through educational or multi-pronged approaches, the results from this review do not show a definitive path to lead poisoning prevention or reduction through education. Despite this, some interventions did show some measures of success in reducing BLL in children, and should be used when more robust options are not possible.

The evidence supporting the use of educational interventions, or how such interventions could be designed to maximize impact is not well-established. Further research on how to design and implement educational and other interventions for effective reduction childhood lead poisoning is needed. Research that examines the most effective and efficient way to disseminate education may be useful, such as in-home education vs. training health care providers to educate within the clinical setting. Due to ethical challenges, conducting a definitive study on the effectiveness of educational interventions for lead poisoning reduction is challenging (if one were to ‘control’ for the amount of lead a child is exposed to determine if education is effective, the only ethical control is to eliminate all known sources lead for all children in all study groups). Future research on this issue must take these challenges into consideration when trying to find a rigorous yet ethical study design.

The studies included in the review spanned over five decades, from 1965, and yet a consistent, improved, or reliable educational intervention did not develop. While this may reflect lack of a coordinated approach to improving and standardizing tested educational interventions, it likely is a stronger reflection the urgency of the need for thorough remediation of lead hazards, rather than reliance on educational lead poisoning prevention interventions. Children, including a disproportionate number of children of color or children from lower socio-economic backgrounds, continue to be poisoned in their own homes through exposure to lead. This poisoning and its lifelong impacts are completely preventable. Although prevention efforts such as full remediation come at a cost, it is imperative that they come. Policies and funding to support abatement of lead hazards are needed to fully remove this scourge from the environment and the bodies of children.