DISCLAIMER
The Lead Exposure Risk Index and its associated data are meant for broad planning purposes only. This tool was created to target outreach and education efforts in high risk zip codes, based on the relative risk of childhood lead exposure across all zip codes with households in NV. A residential zip code LERI score is not a substitute for a clinical risk assessment using a childhood lead risk questionnaire as nationwide datasets and risk factor weights used as inputs for the LERI may not be representative of local data. Please keep in mind that historical and current lead hazards in surrounding areas can also increase the risk of lead exposure to your patients, regardless of the LERI score of their residential zip code.
Methodology
The Nevada Lead Exposure Risk Index (LERI) was developed at the ZIP code level using the following methodology:
Initially, we identified the key variables to incorporate into the index. To inform this selection process, we referred to the work of Jacob and colleagues (Jacobs, et al., 2002), who conducted an extensive study estimating the prevalence of significant LBP hazards in housing units across the country.
We extracted household data by ZIP code pertaining to six critical characteristics from the American Community Survey, specifically focusing on:
- Age of housing
- Housing unit type
- Occupant status
- Household income
- Race
- Ethnicity
These variables were chosen based on their inclusion in Jacob et al.’s study, and we obtained the percentages of housing units with lead hazards from their research:
| Characteristic | Percent with lead hazards |
|---|---|
| Age of housing | |
| After 1980 | 3% |
| 1960-1979 | 8% |
| 1940-1959 | 43% |
| Before 1940 | 68% |
| Housing unit type | |
| Single family | 26% |
| Multifamily | 19% |
| Occupant status | |
| Owner occupied | 23% |
| Renter occupied | 30% |
| Household income | |
| <$30,000/year | 35% |
| ≥$30,000/year | 19% |
| Race | |
| White | 25% |
| African American | 29% |
| Other | 23% |
| Ethnicity | |
| Hispanic/Latino | 32% |
| Non-Hispanic/Latino | 24% |
We made slight adjustments to the first two age of housing categories, “after 1980” and “1960-1979“, to align them with the available categories in our household data. The subsequent step involved calculating , the weighted risk score for each household characteristic, using the following formulas:
- Age of housing
[latex]R_{AH} = \frac{{HH}_{80 +} \times 0.03 + {HH}_{60 – 79} \times 0.08 + {HH}_{40 – 59} \times 0.43 + {HH}_{40 -} \times 0.68}{Total\ Households}[/latex]
- Housing unit type
[latex]R_{TP} = \frac{{HH}_{single} \times 0.26 + {HH}_{multi} \times 0.19}{Total\ Households}[/latex]
- Occupant status
[latex]R_{OS} = \frac{{HH}_{owner} \times 0.23 + {HH}_{renter} \times 0.30}{Total\ Households}[/latex]
- Household income
[latex]R_{HI} = \frac{{HH}_{greater_than_30k} \times 0.35 + {HH}_{at_or_under_30k} \times 0.19}{Total\ Households}[/latex]
- Race
[latex]R_{RA} = \frac{{HH}_{White} \times 0.25 + {HH}_{AA} \times 0.29 + {HH}_{Other} \times 0.23}{Total\ Households}[/latex]
- Ethnicity
[latex]R_{ET} = \frac{{HH}_{hisp} \times 0.32 + {HH}_{nonhisp} \times 0.24}{Total\ Households}[/latex]
, where [latex]{HH}_i[/latex] indicates the number of households of a specific household characteristic.
The third step involved formulating the LERI, recognizing that it constitutes a weighted summation of the aforementioned weighted risk scores. Furthermore, we incorporated the percentage of children ages 6 and under years ([latex]Child6[/latex]) and the poverty rate ([latex]Poverty[/latex]) into the LERI formula.
As a result, the LERI formula appears as follows:
[latex]LERI = w_{AH} \times R_{AH} + w_{TP} \times R_{TP} + w_{OS} \times R_{OS} + w_{HI} \times R_{HI} + w_{RA} \times R_{RA} + w_{ET} \times R_{ET} + w_{Child6} \times Child6 + w_{Poverty} \times Poverty[/latex]
, where [latex]w_{i}[/latex] indicates the risk weight for a given risk score [latex]R_{i}[/latex].
Nonetheless, because of the lacking theoretical support to estimate the eight weights ([latex](w_{AH},\ w_{TP},\ w_{OS},\ w_{HI},\ w_{RA},\ w_{ET},\ w_{Child6},\ w_{Poverty}), the fourth step applied factor analysis to reconstruct the formula using a reduced number of factors that could still adequately represent the original eight variables. The updated formula is expressed as:[/latex]
[latex]LERI = \sum_{i = 1}^{p}{w_{p} \times FS}_{p} + c[/latex]
where [latex]p[/latex] is less than eight, [latex]w_{p}[/latex] is the weight of the pth factor score [latex]{FS}_{p}[/latex], and c is a constant used to scale the LERI, ensuring that it remains strictly positive.
We identified the first four factors, accounting for 31.49%, 21.44%, 13.34%, and 12.18% of variance explained. This collective explanation of 78.44% of the variance from the original eight variables demonstrates the effectiveness of these factors in capturing the underlying patterns.
The fifth step was to compute the weights of the four selected factors. We rescaled their proportions of variance [latex]{Prop}_{x}[/latex] explained by [latex]w_{j} = \frac{{Prop}{j}}{\sum{i = 1}^{2}{Prop}_{i}}[/latex] to make the total weights equal to 1.
The four weights were computed as:
[latex]w_{j} = \frac{{Prop}_{j}}{\sum_{i = 1}^{2}{Prop}_{i}}[/latex]
[latex]w_{1} = \frac{0.3149}{0.3149 + 0.2144 + 0.1334 + 0.1218} = 0.40[/latex]
[latex]w_{2} = \frac{0.2144}{0.3149 + 0.2144 + 0.1334 + 0.1218} = 0.27[/latex]
[latex]w_{3} = \frac{0.1334}{0.3149 + 0.2144 + 0.1334 + 0.1218} = 0.17[/latex]
[latex]w_{4} = \frac{0.1218}{0.3149 + 0.2144 + 0.1334 + 0.1218} = 0.16[/latex]
Therefore, the final index score can be formulated as:
[latex]LERI = 0.40 \times {FS}_{1} + 0.27 \times {FS}_{2} + 0.17 \times {FS}_{3} + 0.16 \times {FS}_{4} + 1.12[/latex]
The factor analysis was exclusively applied to ZIP codes with households. ZIP codes without households are postal routes that correspond to address groups or delivery routes and are not a representation of physical boundaries, buildings, or populations. Consequently, a total of 15 ZIP codes without households were excluded from the analysis.
Lastly, we categorized the calculated LERIs into deciles, with each decile representing 10% of the remaining 175 ZIP codes in Nevada:
| Decile | LERI |
| 1st | 0.10-1.33 |
| 2nd | 1.33-1.75 |
| 3rd | 1.75-2.09 |
| 4th | 2.09-2.31 |
| 5th | 2.31-2.46 |
| 6th | 2.46-2.52 |
| 7th | 2.52-2.56 |
| 8th | 2.56-2.62 |
| 9th | 2.62-2.68 |
| 10th | 2.68-2.95 |
Geographic Areas of Priority
ZIP codes in deciles 1, 2, and 3 are considered low-risk.
ZIP codes in deciles 4, 5, 6, and 7 are considered medium-risk.
Lastly, ZIP codes in deciles 8, 9, and 10 are considered as areas at high-risk for lead exposure and are NvCLPPP’s areas of priority.
| Decile Range | LERI Range | Ranking |
| 1st-3rd | 0.10-2.09 | Low-risk |
| 4th-7th | 2.09-2.56 | Medium-risk |
| 8th-10th | 2.56-2.95 | High-risk |
Figure 1 highlights high-risk ZIP codes for the state.
While blood lead surveillance data was not incorporated at this time due to low screening rates, the NvCLPPP plans to improve the collection of epidemiologic data to be able to include race and ethnicity of BLL above the BLRV and other BLL data in future surveillance maps.