C L P E R
A Not for Profit Community Based OrganizationCHILDHOOD LEAD POISONING
EMERGENCY RESPONSE
PO Box 1209 ¥ Maplewood, NJ 07040
Phone: (973) 763-9566 ¥Fax: (973) 763-5551
E-mail:
WWW:
Putative Declines in Elevated Blood Lead Levels: Real or Illusory?
A Review of NJ DHSS Childhood Lead Poisoning in New Jersey, Annual Report Fiscal Year 2001
©Myles OÕMalley, MA
( revised 2/05/02)
The report ÒChildhood Lead Poisoning in New Jersey, Annual Report Fiscal Year 2001,Ó (the Report 2001) is the fourth to be issued as required by PL 1995, Chapter 338. The information in this and last yearÕs Report 2000 has improved significantly. It should nonetheless be subjected to interpretation and closer analysis than, in some instances, provided by NJDHSS. Although the Report 2001 contains much information that requires reflection and research[1], this review will focus on the putative declines in elevated blood lead levels and simultaneous increase in rates of lead screening. The issue is pivotal because New Jersey has a new administration and a severe budget deficit that by statute cannot be tolerated. It is imperative that all constituencies, including the State, have a fundamentally common perception of the extent, breadth and location of this disease of childhood poverty. Policy recommendations will be made.
DECLINE IN REPORTED ELEVATED BLOOD LEAD LEVELS (EBLLS): REAL OR ILLUSORY?
The Report 2001 is the latest in a series that show a steady and dramatic decrease in the number of reported elevated blood lead tests in New Jersey. The Report 2001 shows that 3.8 % of children tested had elevated blood lead levels; i.e. levels => 10mcg/dL down from 5% in last yearÕs report. The actual number of children with elevated blood lead levels (EBLLs) declined 18% from 6,847 to 5616. Even more startling is the fact that the number of children with EBLLS => 20mcg/dL (the level at which environmental intervention is required by law) declined 27% to 947 down from 1309 in last yearÕs report. The Report also for the first time supplies blood lead test results dating back to 1994 for children with blood lead levels => 20mcg/dL. In 1994, there were 4757 children tested with blood lead levels of =>20 mcg/dL. The report shows a steady annual decline to the 2001 level of 947. This represents an 80% decline in this blood lead category. How are these declines explained?
There is no doubt that the Report 2001 reflects the blood lead levels of the children tested. Do these declines reflect actual declines in the populations most at risk for elevated blood lead levels? Although the Report does not categorically state that this is the case, it and the press release upon which it is based strongly suggest that that conclusion is plausible while advancing no explanations for the declines. The press release (December 17, 2001) is self-congratulatory without qualification: Ò ÔI am pleased to see that New Jersey is making steady, consistent progress in identifying children with lead poisoning,Õ said Acting Governor Donald T. DiFrancesco.Ó
From the Report 2001:
Trends in testing and results
These results represent the continuation of a long-term trend of decreasing numbers of children identified with elevated blood lead. The DHSS established the reporting level for elevated blood lead at 20 mcg/dL in 1993. State FY 1994 was the first full State Fiscal Year that this reporting level was in effect. Every year since then, the number of children reported with blood lead at this level or greater has declined É. In the absence of reporting of all test results, the DHSS was not able to determine if the reduction between FY 1994 and FY 2000 was due to a real decline in elevated blood lead among children, or due to a reduction in the number of children tested for blood lead. [As a consequence of transfer from Fee for Service to Managed Care, comment added]. The fact that the number of children with elevated blood lead continued to decline from FY 2000 to FY 2001, even as the number of children tested increased, may indicate that the decrease in reported children with elevated blood lead reflects a real decline in elevated blood lead in children in New Jersey throughout this period. This finding would be consistent with reductions in children with elevated blood lead reported to CDC by other states, as well as the reduction in the average blood lead level in children in national surveys.
[Reference: US Centers for Disease Control and Prevention, ÔBlood Lead Levels in Young Children Ð United States and Selected States, 1996-1999,Ó Morbidity and Mortality Weekly Report, December 22, 2000, 49 (50): 1133-7.].[2]
The decline in childhood lead screening referred to above harkens back to the NJ 1999 Report which described a 75% reduction of lead screenings from 37,697 in FY 1995 to 9,586 in FY 1999. In the same time period, there was a 62 % reduction in blood lead levels => 20mcg/dL.[3] Ironically, this reduction occurred despite the passage of the universal lead screening law in 1996 and its implementation in 1998. The decline also corresponded precisely with the large-scale enrollment of Medicaid children into Managed Care. The inference was that HMO procedural obstacles caused the decline in screening and consequent reported decline in elevated blood lead levels. The current report holds that inference suspect because screening rates, though remaining low as a percent of all those children who should have been screened, nonetheless increased in the aggregate in FY 2000 to 137,536, the first year a DHSS requirement that laboratories report all blood lead tests, and then increased by 11,697 (8.5%) for FY 2001. More children screened, dramatic reductions in numbers of lead poisoned children. How can this not be Òsteady progress?Ó
REPORT 2001 OFFERS NO EXPLANATIONS FOR THE REPORTED DECLINES
In the first instance, a decline of 80% in children with blood lead levels => 20mcg/dL from 1994 to 2001 is, without explanation[4], preposterous, and most certainly does not comport with CDC data reported in MMWR for December 22, 2000 cited by the Report 2001 above. As the CDC report makes clear, the 80% declines in childrenÕs blood lead levels as reported in the Third National Health and Examination Survey (NHANES III, 1991-1994) commencing in the late 1970Õs, was primarily the result of de-leading gasoline, the consequences of which were complete by the early 1990Õs. NJÕs putative 80% decline commenced in 1995 and had nothing to do with the de-leading of gasoline.
EBLL DECLINES IN NEW JERSEY DO NOT COMPORT WELL WITH CDC DATA
The Report 2001 invites comparison between NJ data and the MMWR for December 22, 2000. The MMWR paper reflects blood lead tests reported to state surveillance programs for three years 1996-1998. Among the 19 states there was considerable variation, from 2.7 to 14.9, in the percentages of children tested and found to have blood lead levels => 10mcg/dL. Within individual states there was also considerable variation from county to county. That range was from 0.5% to 27.3 % lending further credence to the proposition that childhood lead poisoning is localized and concentrated. Does New Jersey data follow the trend indicated by the MMWR paper?
The following table is reprinted from the MMWR paper:
Table 1. Percentage of children tested aged <6 years with elevated blood levels (EBLLs), by year Ð Selected states, 1996-1998.
% Children with elevated BLLs (mcg/dl)*
YearNo. Tested=>10+>15=>20
1996 / 1,220,596 / 10.5% / 3.9% / 1.9%1997 / 1,183,506 / 8.6% / 3.2% / 1.5%
1998 / 1,256,907 / 7.6% / 2.7% / 1.2%
* Alabama, Colorado, Connecticut, Iowa, Maine, Massachusetts, Michigan, Minnesota, Montana, New Hampshire, New York, North Carolina, Ohio, Oklahoma, Utah, Vermont, Washington, Wisconsin, and Wyoming.
For three years, from 1996 to 1998 there was a decline of 27% in elevated BLLs => 10mcg/dL in the 19 states. Using a three-year period for NJ data, Annual reports showed a rate of 15.9% for FY 1999, 5% for FY 2000 and 3.8 % for FY 2001. This represents a decline of 76% from FY 1999 to FY 2001. This is not in keeping with the MMWR decline of 26%. The 2-year decline from FY 2000 to FY 2001 is also not in line with the 19 state trends for 1997 and 1998. New JerseyÕs decline is 24 % as opposed to 7% for the nineteen states.
A comparison between the 19 state decline in elevated BLLs =>20mcg/dL and NJÕs data similarly does not show comparability. From the Table above, the decline in this category over the three-year period is 37%. The NJ Report 1999 showed 3% of children tested with BLLs => 20mcg/dl, the Report 2000 showed 1%, and the Report 2001 showed 0.6%. This is a decline of 80% over the three-year period as compared to the 19 state declines of 37%. From a two-year period, FY 2000 to FY 2001, the NJ decline was 40% as compared to the 19 state declines of 20%.
Across all categories of EBLLs New JerseyÕs declines for similar time frames are at least twice as large as the 19 state averages and in some instances three times as large. It is also important to note that the MMWR report does not account definitively for the declines in elevated BLLS. Consequently, the explanations hypothesized in this Review of New JerseyÕs declines may also apply nationally.
What follows are hypotheses that explain the NJ blood lead level declines. They require further research. But it is hoped that the discussion will foster real debate.
CHILHOOD LEAD POISONING IS LOCALIZED AND CONCENTRATED
CDC data has made it clear that childhood lead poisoning is very much a targeted or localized phenomenon. According to NHANES III, Phase 2 data, the geometric mean of BLLs for children 1-5 years was 2.7 mcg/dL. That same study showed that low-income children living in older housing had more than a thirty-fold greater prevalence of BLLs => 10 mcg/dL. Because EBL children are so concentrated, CDC has recently issued its recommendations for targeted screening of Medicaid populations.[5]
Analysis of New JerseyÕs decline in reported EBLLs is governed from beginning to end by the fact that childhood lead poisoning strikes targeted populations. From this standpoint, we look closely at the 1994-1995 data that showed the high numbers of children with EBLLs. Next we ask whether there have occurred any significant social changes in high-risk populations or their environments that would affect the likelihood of being screened or the levels of lead in their blood. The following graph is reprinted from the Report 2001:
In 1994 and 1995, the extraordinarily high number of children with BLLs => 20 mcg/dL was the result of highly targeted system that tested 37,697 children from July 1, 1994 through June 30, 1995.[6] Many of these children (31, 910) were tested by Local Health Departments in their Child Health Conferences (well-child clinics), the remainder presumably in physician offices as part of fee for service health care delivery. Local Health Departments tested these children if they had Medicaid coverage or were uninsured with the family income less than 250% of the Federal Poverty Level (FPL). The poor and ethnic minorities were highly represented. We know that these high numbers of EBLLs also reflected a NJ screening rate of 39% for Medicaid recipients under Early and Periodic Screening, Diagnosis and Testing (EPSDT), a high rate relative to most other states where the average was 20%.[7]
THE REPORT 2001 STATEMENT THAT EBLLS HAVE BEEN IN DECLINE FROM 1994 TO THE PRESENT CANNOT BE SUPPORTED!
According to the Report 1999, there was a decline of children screened for lead poisoning from the 37,697 in FY 95 to 9586 in FY 1999, a 75% decline. The number of children with BLLs =>20mcg/dL in FY 1995 was 4187 and in FY 1999 was 1602 for a decline of 62%. Looking at the data in this time frame from FY 1995 to FY 1999, one would have to conclude that the decline in EBLLS => 20 mcg/dL was the result of declines in the number of screenings. Since the rate of EBLLs per screening actually increased, one would also have to conclude that those children screened remained highly targeted. The Report 1999 points out that the decline in screenings exactly coincides with the migration from fee for service to Managed Care. This is correct and explains, at least in part, the decline in screenings and EBLLs detected. It overlooks the equally important issue of the rate of EBLLs per number of children screened.
In both groups selection criteria remained the same; i.e., many children were screened by Local Health Departments in their Child Health Conferences using the same selection criteria as stated above (Medicaid Recipient or uninsured with income less than 250% of FPL) and others were screened in other settings with blood lead levels analyzed by private, licensed labs. Thus the children screened, though small in number, continued to represent a targeted population. The disproportion between the 75% decline in the number of children screened in this time frame and the decline in the number of children with EBLLs, 62%, could be interpreted as an increase of 13 % in EBLLs =>20-mcg/dL s or as the result of the 1999 group for some reason being more targeted. Either way, the rate of children with EBLLs -> 20 mcg/dL actually increases. The Report 1999 accepts the reported declines of EBLLs at face value and proposes the migration to Managed Care as the cause, leaving open the question whether they represent real declines in the population of children.
The data also does not support a decline in EBLLs when one looks only at children screened by Local Health Departments for FY 1998 and FY 1999. The percent of children with EBLLs => 10mcg/dL for 1998 was 14.5% and for 1 999 was 15.89% for an increase of 9.6%. Similarly, for 1998 the percent of children with EBLLs =>20mcg/dL was 2.8 and for 1999 was 3.7 % for an annual increase of 32%.[8]
From FY 1994 through FY 1999, the data simply does not support even the speculation that the decline in reported EBLLS represented real declines for the population of children as a whole. Indeed, the data points in an opposite direction.
The data for Reports 2000 and 2001 present a somewhat different analytical challenge, since they reflect a change in NJDHSS policy that required, as of July 1, 1999, that all blood lead tests be reported to the state and not just levels => 20mcg/dL. The number of reported screenings for FY 2000 was 137,536 and for FY 2001 was 149,233. As stated previously, despite the increased screenings the decline in EBLLs continued dramatically (an 18% decline in children with EBLLs => 10 mcg/dL and a 27% decline in children with EBLLs => 20mcg/dL.) The answer to the question ÒDo these reported declines represent actual declines in EBLLs among high-risk children?Ó will revolve around the extent to which the screenings remained targeted. Were there changes to the social context of the high-risk population of children or their environment that could have affected whether high-risk children were screened for lead poisoning?
WELFARE REFORM, STATE CHILD HEALTH INSURANCE PROGRAM (SCHIP) AND EBLL CHILDREN
What follows develops the hypothesis that on the one hand, the population of high-risk children actually screened from July 1, 1999 through June 30, 2001 declined dramatically as the result of the implementation of Temporary Assistance to Needy Families (TANF) and its attendant decline in Medicaid rolls. On the other hand, the decline was masked by the passage of NJÕs Universal Screening Law and the increased enrollment of children in NJ KidCare, both of which would have the effect of increasing screenings while not necessarily reaching targeted populations. According to a GAO report already cited, 80% of EBLL children are Medicaid eligible. Thus changes to the demographics of Medicaid enrollment deserve scrutiny in the assessment of the significance of reported declines in EBLLs.
For NJ DHSS Report 2000, the first year when all blood lead tests in the State were reported, not only is there a decline in the number of EBLLs across all categories but the rates per children screened also declines. New Jersey HCFA-416 report for FY 1999 indicated a total of 20,056 lead screenings for children one through five years of age. For this age group, the rate of screening was 14.5 % (141,075 eligible for lead screening).[9] In this period, NJDHSS reported 1603 EBLLs =>20mcg/dL for all children tested. NJ HCFA ReportÐ416 for FY 2000 saw a marked increase in EPSDT lead screenings to 36,810 for ages 1 through 5, representing a screening rate of 25% (145,788 eligible for lead screening). Our NJDHSS Report 2000 indicates 1309 children with EBLLS =>20 mcg/dL out of a total of 135,572 children tested.
1999 / 135,572 children tested / 1603 found with EBLLS =>20 mcg/dL / Rate = 14.5%2000 / 145,788 children tested / 1309 found with EBLLS =>20mcg/dL / Rate = 25%
WhatÕs wrong with this picture?
The Trend Continues: As already noted, the Report 2001 cites an increase of 11,697 screened in FY 2001 for a total of 149,233 identifying 947 with EBLLS of =>20 mcg/dL. The easy interpretation of the data is that there actually has been a real decline not in just reported EBLLs but in EBLLs statewide including high-risk populations. Because it affects those least able to fend for themselves, our children, this is a conclusion that cannot be accepted without explanation and validation. Other hypotheses must be advanced and tested. Since childhood lead poisoning and Medicaid eligibility are so closely correlated, one is compelled to look at changes to the Medicaid population for answers.[10]
TANF AND MEDICAID
From 1994 through 1996, a significant percentage of the lead screened population was enrolled in Welfare, Aid to Families with Dependent Children (AFDC). Welfare enrollment carried with it automatic participation in Medicaid. With the passage of the Personal Responsibility and Work Opportunity Reconciliation Act of 1996 (PRWORA), TANF, the program that replaced AFDC, and Medicaid were de-linked. TANF recipients had to formally apply for Medicaid and despite provisions in welfare reform law designed to assure continued and even extended Medicaid coverage[11], TANF resulted in significant numbers of Medicaid recipients losing coverage. Denial of Medicaid benefits associated with Welfare de-linking was deemed of sufficient magnitude to prompt HCFAÕs Director to require states to determine the extent of their individual problems and to take measures to re-instate coverage, as required by law, that was improperly terminated.[12]
Nationally, for adult women who leave welfare slightly less than half are enrolled in Medicaid six months after leaving. After a year or more off welfare this percentage decreases to less than a quarter and about one half have no health insurance at all. For the children of these women, Medicaid coverage remains strong six months after leaving welfare at 75% but after a year or more less than half of these children have Medicaid coverage and almost one third have no health insurance at all.[13] This increasing decline over time is a function of the terms under which Medicaid can be retained, including time limits for transitional Medicaid coverage, face-to-face interviews, complicated application processes requiring difficult to obtain information, and inadequately trained caseworkers.[14]