RATIONALE FOR THE COMPARISON OF THE
TCE SUBREGISTRY DATA WITH NATIONAL DATA
This section includes comparisons of TCE Subregistry data with data from national surveys. These comparisons are consistent with Registry objectives and goals as stated in the Polices and Procedures Manual (1), which are to provide a preliminary assessment of the extent to which TCE Subregistry members have an excess of adverse health conditions and to generaterather than testhypotheses about TCE exposure and health outcomes.
In addition to a comparison of Subregistry health data with national health data norms, this section includes a comparison of registrant demographic and smoking data with national data. The comparison of demographic characteristics indicates the extent to which TCE Subregistry members are similar to the general population. The comparison of smoking rates provides an assessment of the extent to which the smoking rate among registrants is similar to smoking rates found in the general population. The latter two comparisons are important because both demographic characteristics and smoking are known to be correlated with or are probable causes of adverse health conditions.
NATIONAL COMPARISON POPULATIONS
TCE Subregistry data were compared with data obtained from the National Health Interview Survey (NHIS)(14) and the National Household Survey on Drug Abuse (NHSDA)(15). The NHIS data were used in the comparisons of demographic and health data. The NHSDA data were used to compare TCE Subregistry smoking rates with the national rates. The NHSDA data were used for comparison purposes because the 1989 NHIS survey did not include questions about cigarette smoking for the subpopulation used in the data analyses presented in this report.
Selection of NHIS
Subsets of the NHIS data were selected as the comparison data for the National Exposure Registry components. The NHIS is an appropriate comparison population because it is a subset of the residential, noninstitutionalized U.S. population, the population of interest for comparisons of the health status of the National Exposure Registry members. As of 1985, a stratified, multistage cluster sample design was used in the NHIS to obtain a representative sample of the target population; this information was used to create representative national norms. The NHIS, similar to the Registry, is composed of self-reported data that were obtained using face-to-face interviews.
Because of the similarity of the data collection instrument and methods used by the NHIS and the National Exposure Registry, the NHIS data were appropriate for the calculation of selected prevalence and incidence statistics and could be used for exploratory comparison with Registry data for health outcomes. The weighting factors (16) provided by the National Center for Health Statistics (NCHS) were applied when using the data; otherwise, because of the NHIS sample design, the estimates would have been biased. The 1989 NHIS file used for selected comparisons in this report included 116,929 respondents.
The NHIS was selected for comparison with the TCE Subregistry for the same reason that NHIS was selected for comparison with the National Exposure Registry in general. As discussed in Section 4 of this report, about 90% of the TCE Subregistry sample is located primarily in the East North Central United States (in Illinois, Indiana, and Michigan), with the remainder located throughout the other regions. Despite the regionality of the TCE Subregistry when compared with the NHIS, ATSDR's review of the regional rates for selected outcomes found no definitive evidence indicating that the overall health status of midwesterners differed significantly from that of the general U.S. population. Rates for selected reported chronic conditions for the geographic regions Northeast, Midwest, South, and West (included all races) are listed in an NCHS publication (14). Of the 65 conditions reported, the rates for 14 were highest in the Midwest region. Of these, bladder disorders are of particular interest when discussing the TCE Subregistry results. Bladder disorders were reported at a rate of 1.7 cases per 100 people for the Midwest (includes all races) and at a rate of 1.6 for the next highest region; of the registrants (white only) the males had a rate of 1.7 cases per 100 people, the females 5.5. For selected diseases that were reported most often by the general population in the Midwest region, the registrants' rates were higher than the regional rates. Therefore, differences between the TCE Subregistry file and the NHIS file were not expected to be the result of regional differences.
Selection of NHSDA
The NHIS 1989 file, the version used for comparison, did not contain information on cigarette smoking, except for a small subpopulation of women of childbearing age. For that reason, this report uses published cigarette use estimates from the 1990 version of the NHSDA (15) for comparison with the TCE Subregistry smoking data. The 1990 NHSDA is the 10th of a series of annual national surveys intended to measure the prevalence of drug use among Americans 12 years of age and older. The NHSDA used a probability sampling design that involved sampling households in the continental United States from 100 primary sampling units. Within the households selected, subjects were randomly selected. The NHSDA contains self-reported data that were obtained via face-to-face interviews. These interviews involved the use of self-administered answer sheets to minimize respondent concerns about confidentiality or anonymity for some sensitive questions. The 1990 NHSDA file included 9,259 respondents.
METHODS OF DESCRIPTIVE VARIABLE COMPARISONS
Demographic Characteristics
The NHIS and TCE Subregistry samples were compared in terms of four demographic characteristics: sex, age, race, and educational level. Each of these variables, along with cigarette smoking, is a potential correlate of health status.
Sex
The distribution of the male-female ratio was assessed on an age-specific basis. The proportion of males and females in each age category was based on the NHIS data and compared with the corresponding proportions in the TCE Subregistry. Each age-specific proportion in the TCE Subregistry was compared with the corresponding proportion in the NHIS by testing that the binomial proportion was equal to a specified theoretical value. No significant differences were found between the two files for this variable.
Age
The descriptive comparisons of age used a 10-category measure. The regression analyses in this section involved a regrouping of age categories. An eight-category measure of age (combining the lower two and upper two groups) was used because of the sparsity of positive reports in some of the age strata.
It should be noted that because the health outcome analyses involved summarizing age- and sex-specific comparisons rather than analyzing age-adjusted summaries, whether the age distribution of the NHIS file matched the age distribution of the TCE Subregistry file was not directly relevant unless numbers differed within the age groups.
Race
Race is an established correlate of socioeconomic status (17) and health status (18). National data indicate that nonwhites have lower rates for cigarette smoking (15). For these reasons, race is a potential control variable for the comparisons of health status and smoking rates. However, as will be discussed in a subsequent section, there were too few nonwhites in the TCE Subregistry to use race as a variable; all analyses were restricted to registrants responding "white" to the race question.
Education Level
For education level (the highest level attained as reported by a respondent), the descriptive analyses included comparisons in which education level was measured as a seven-category ordinal variable (that is, 0 through 5 years, 6 through 8 years, 9 through 11 years, 12 years or the equivalent of a high school diploma, 13 through 15 years or some college, 16 years or the equivalent of a college degree, and 17 or more years).
For the regression analyses, education attainment was collapsed into three groups: 0 through 11 years, 12 years, and 13 or more years. Because education is difficult to interpret as a surrogate for socioeconomic status for school-aged children, analyses involving adjustments for education were restricted to adults (19 years of age or older).
Cigarette Smoking
Rates for current and past smoking behavior were compared across sex, age, and educational attainment categories. A current smoker ("current rate") was defined as anyone who reported being a smoker at the time of the interview, and who had smoked at least 100 cigarettes in his or her lifetime. Past smoking behavior ("ever rates") was assessed by calculating the rates for people who had ever smoked at least 100 cigarettes during their lifetime. People who had ever smoked included both current and ex-smokers.
METHODS OF COMPARING HEALTH OUTCOMES
Question Comparability
TCE Subregistry and NHIS data were compared for health conditions reported by respondents. Such comparisons were preceded by an assessment of the comparability of NHIS and TCE Subregistry health condition questions. The questions about health conditions in these two surveys differed in three respects: restrictions on the source of diagnosis; the time frame of occurrence or treatment; and, in some cases, the wording of the health condition. A discussion of each potential source of variation in health condition questions follows. The NHIS health-related questions are in Appendix B-1 and the TCE Subregistry health-related questions are in Appendix B-2.
Source of Diagnosis
TCE Subregistry questions about health conditions specified that the source of diagnosis must be a "physician or other medical provider." This qualification was intended to minimize self-diagnoses or the biased reporting of health problems by registrants, since they might have a greater awareness of health because of their known exposure and publicity related to the exposure. The NHIS questions did not include any type of qualification concerning the source of diagnosis. Therefore, if all other factors were equal or the same, an increased reporting by NHIS respondents when compared with the TCE registrants might be expected.
Time Frame
TCE Subregistry baseline questions about health conditions asked about diagnoses of or treatment for conditions from the point of birth through the date of the interview ("Has a physician or other medical provider ever told you/SUBJECT that you/he/she/ had or treated you/SUBJECT for CONDITION?"). Only one time frame was addressed: ever had (subject's lifetime). Respondents who reported "yes" to this question were also asked whether the subject was ever treated for the condition, when the subject was first treated for the condition, and whether the subject was currently being treated for the condition.
The NHIS questionnaire included questions that focused on three time frames--ever had the condition, had the condition within the last 12 months, or currently had the condition. With the exception of heart disease, only one time frame was used to create a reponse rate for any given health condition. The NHIS data file heart disease rate reflected a composite of responses to heart-related questions that were asked in both the "ever" and "12-month" time frames. The NHIS questionnaire asked whether respondents had ever had the heart conditions rheumatic, congenital, or coronary heart disease; angina pectoris; myocardial infarction; or any other heart attack; in the last 12 months time frame for damaged heart valve, tachycardia or rapid heart, heart murmur, or other heart trouble. In addition, for some heart-related questions, a positive response was discarded if the respondent did not answer positively to one or more other selected questionnaire items (16). A comparable heart response rate could not be created for the NER file and, therefore, a comparison with national norms could not be made for the heart condition variable.
For the other health conditions, the time frames were standardized to make the NHIS and TCE Subregistry rates directly comparable. Table 5-1 provides a comparison of NHIS and TCE Subregistry questions in terms of the time frame for each health condition. One NHIS health condition question, the effects of a stroke, was asked and rate calculated in the context "have you ever had." The questions and time frames for the subregistry and NHIS matched on this condition.
Table 5-1.--Comparison of time frames for health condition questions.
TCE Subregistry Conversion from "ever had" to |
NHIS Version: "ever had" |
NHIS Version: "in the past 12 months" |
NHIS Version: "now have" |
"Ever had" | Stroke | ||
"in the past 12 months"
("ever had" and "currently have" and/or date of 1st treatment within past 12 months) |
Cancer, rash, anemia, kidney
disease, urinary tract disorders, ulcer, liver prob- lems, asthma, respiratory problems and allergies, diabetes, arthritis, hyper- tension |
||
"now have"
("ever had" and "currently have") |
Speech impairment,
hearing impairment, mental retardation |
Eleven of the NHIS questions were asked in the time frame "in the past 12 months." For hypertension, the NHIS 12-month response rate was calculated using the "ever had" positive responses; however, the positive response was retained in the file only if the respondent also answered positively to one or more of nine other selected questionnaire items (16). This additional restriction might have reduced the NHIS response rate for this condition. For these 12 health conditions (see Table 5-1), the TCE Subregistry time frames for comparison were adjusted. In the TCE Subregistry, a health condition was defined as occurring "in the past 12 months" if (1) the reported date of first treatment was within the 12 months preceding the interview or (2) the subject was receiving treatment at the time of the baseline interview. This adjustment could have resulted, however, in an underestimation of these 12 conditions for TCE Subregistry data for the following reason. A year or more prior to the baseline interview, a registrant might have been told that he or she had (or was treated for) one of these 12 health conditions, but was not being treated at the time of the interview. If so, such a registrant would not have been included in the rates for these 12 health conditions.
There were three health conditions in the NHIS questionnaire that were queried in the time frame "do you now have." These conditions were speech impairment, hearing impairment, and mental retardation. The time frame for the comparable TCE Subregistry health conditions was adjusted by counting only registrants who reported that they were "currently receiving treatment" for one of these three conditions. Again, if all other factors were equal or the same, an increased reporting by the NHIS respondents when compared with the TCE registrants would have been expected.
Health Conditions
TCE Subregistry and NHIS questions were also compared in terms of the phrasing of health conditions. As Table 5-2 indicates, some health conditions matched exactly, others did not. An ATSDR panel of scientists and physicians determined matches for the TCE Subregistry health conditions and specific NHIS conditions (ICD-9 codes (19), or NHIS condition recodes (16)).
The nine health conditions in Class A of Table 5-2 either matched exactly or the TCE Subregistry version was inclusive of the NHIS version. That is, the NHIS wording of the health condition and the NHIS classification of the condition in the recodes were the same as or paralleled closely the corresponding TCE Subregistry item. As Table 5-2 indicates, nine health conditions were in Class A. Class B included seven health conditions that did not match as closely, but were considered to be sufficiently similar for the purposes of the NHIS and TCE Subregistry comparisons.
In addition to the heart disease outcome, for six other health conditions on the TCE Subregistry questionnaire there were no parallel items in the NHIS questionnaire. These conditions pertained to symptoms, including "frequent periods of fatigue or tiredness"; "frequent periods of nausea"; "seizures, tremors, spells, or epilepsy"; "weakness or paralysis or numbness in the arms or legs"; "frequent periods of anxiety, nervousness, or depression"; and "frequent or severe headaches."
Unlike some of the national health surveys (20), environmental studies commonly ask about symptoms as well as health outcomes. Data on these symptoms, while not directly comparable with NHIS data, are important in assessing environmental impact on health and will be useful for comparisons with past and current epidemiologic environmental studies, as well as for future longitudinal studies. Symptom response rates were compared in intrafile exposure groups.
The TCE Subregistry questionnaire was used to record information on all types of cancer via an open-ended question. The NHIS questionnaire, however, was used to obtain direct information on only some types of cancers, including skin, stomach, intestinal, colon, rectal,
Table 5-2.--Comparison of TCE and NHIS health questions.
Q#* | Wording in TCE Survey | NHIS Definition | NHIS
Chronic Recodes† |
ICD-9§ |
Class A¶ | ||||
6 | High blood pressure (hypertension) | Essential hypertension
Hypertensive heart disease Hypertensive renal disease Hypertensive renal and heart disease |
C508 | 401-05 |
8 | Kidney disease | Kidney stones
Kidney infections Other kidney trouble |
C409-11 | 592
590 581-3 593 |
10 | The effects of stroke | Cerebrovascular disease | C509 | 430-38 |
14 | Liver problems | Liver disease, including cirrhosis | C302 | 571-2
573.0, .3-.9 |
15 | Asthma, emphysema, or
chronic bronchitis |
Same | C601-2
609 |
490-1
492 493 |
16 | Other respiratory
allergies or problems such as hay fever |
Hay fever
Allergic rhinitis without asthma |
C603 | 477 |
17 | Diabetes | Same | C403 | 250 |
22 | Hearing impairment | Deaf - both ears
Other hearing impairment |
C203-4 | X05
X06-9 |
25 | Mental retardation | Same | C208 | X19 |
Class B** | ||||
3 | Cancer | Some cancers queried directly;
other ascertained indirectly | 140-203 | |
5 | Skin rashes, eczema, or
other skin allergies |
Psoriasis
Dermatitis Dry (itching) skin |
C112-4 | 696
690-94 698.9 |
7 | Anemia or other blood
disorders |
Anemia of any kind | C508 | 401-05 |
9 | Urinary tract disorders,
including prostate trouble |
Disorders of the bladder (other
than bladder infections) Diseases of prostate |
C413-14 | 594.1
596 600-602 (except 601.4) |
Table 5-2.--Continued.
Q#* | Wording in TCE Survey | NHIS Definition | NHIS Chronic Recodes† | ICD-9§ |
Class B** | ||||
13 | Ulcers, gall
bladder trouble, or stomach or intestinal problems |
Gallbladder stones
Gastric, duodenal, or peptic ulcer Abdominal hernia Gastritis and duodenitis Disease of esophagus Other functional disorders of stomach or digestive system (not indigestion) Enteritis and colitis Spastic colon Diverticula of intestines Other stomach and intestinal disorders (not constipation) |
C301
C303 C303-8 C310-3 C315 |
574
530-7 550-3 555 556 558 560.562 564.1 569 787 |
18 | Arthritis,
rheumatism, or other joint disorders |
Arthritis
Rheumatism Gout Sciatica (and lumbago) Intervertebral disc disorders Bone spur and tendinitis Disorders of bone or cartilage Bursitis |
C101-7
C109 |
711.0, .9
712.8-.9 714-16 720.0 721 729.0 724,.2-.3 722, 726 727.0, .2-.9 730.0-.3, .9 731.0, .2 732-3 |
20 | Speech impairment | Stammering and stuttering
Other speech impairment |
C205-6 | X10
X11 |
No Match | ||||
19 | Rheumatic fever, heart disease, or other heart
problems |
Rheumatic fever
Ischemic heart disease Heart rhythm disorders Congenital heart disease Other select heart diseases |
C501-7 | 390
392-9 410-4 427.0-.6, .8-.9 785.0-.2 745-6 415-7, 420.9 421.0, .9 422.9 423-4 425.0-.5, .9 426, 428 429.0.6, .8-.9 |
*Question in TCE Subregistry questionnaire.
†Chronic Recodes, NHIS, Public Use Data Tape Documentation (18).
§ICD-9 is the International Classification of Diseases, 9th Revision, World Health Organization (WHO) (19).
¶Class A indicates questions match exactly or closely.
**Class B indicates questions are similar.
lung, breast, and prostate cancer; information on other cancers was obtained indirectly by the respondent on hospital stays, doctor visits, and restricted activity. The NHIS question is worded, "In the last 12 months, did anyone in the family have .... cancer?" The time frame restriction
and the possible restriction on types of cancers reported makes this comparison with the TCE registrant data questionable and the interpretation tenuous.
Statistical Analysis of Health Data
The statistical analyses performed treated the NHIS population as a standard population and applied the age- and sex-specific period prevalence and prevalence rates obtained from the NHIS data to the corresponding age- and sex-specific denominators in the TCE Subregistry. The observed age- and sex-specific numerators for the TCE Subregistry were compared with the expected numerators, based on the NHIS rates.
This one-sample approach ignored sampling variability in the NHIS data because of the large size of the NHIS database relative to the TCE Subregistry data file. Given that the primary focus of this report is the TCE Subregistry, treating the NHIS versus TCE comparison as a two-sample problem might have resulted in a dramatic underestimation of the variability associated with the TCE data if any pooled variance estimates were used. For the two-sample problem, an estimation of the standard errors was made before the statistical comparisons were made. For the NHIS data set, this report used the established methods and estimates of rates that the NCHS published (14). In estimating standard errors for the TCE Subregistry data, the assumption was made that the binomial distribution applied.
This report used the person-weights (or "final basic weight"(16)) in calculating the health condition rates for the NHIS data. To allow for the nonequiprobable sampling in the NHIS data, all age- and sex-specific period prevalence and prevalence rates that were derived from the NHIS data were weighted by the appropriate person-weights. These weights reflect the complex sampling method used by the NCHS in the survey design (16). The Primary Sampling Unit codes contained on the NHIS data file allow for partial adjustment for the clustering component of the NHIS survey design, but the adjustment was not used in the analyses for this report. The age- and sex-specific rates were calculated using only those respondents who were queried about each of the conditions of interest. All health outcomes were analyzed in the following manner. Taking the NHIS as a standard population, weighted age- and sex-specific period prevalence and prevalence rates were constructed using the person-weights. These "standard" rates were applied to the corresponding TCE Subregistry denominators to obtain expected counts in each age and sex combination.
Paralleling Poisson regression modeling of standardized mortality ratios, the ratios of the observed-to-expected age- and sex-specific counts were modeled using Poisson regression in the Generalized Linear Interactive Modeling program (21). The Poisson regression approach is described in Breslow and Day (22). Maximum likelihood estimation was used, and likelihood ratio statistics and Wald confidence intervals (when appropriate) were computed using EPICURE software (23). In the Poisson regression analysis, the null model is specified by log(observed) = log(expected) + grand mean. The hypothesis that the rates are the same can be rejected when a Wald confidence interval about the grand mean does not include the value of 1, provided that the null model is adequate.
By adding terms for age and sex effects to this null model, it was possible to detect structure (confounding) in the ratios of observed-to-expected prevalence rates as a function of these variables. To take education into account, each analysis was repeated for adults (aged 19 years and older) using age-, sex-, and education-specific rates.
Statistical Analysis of Cancer Data
For the cancer health outcome, cancer incidence data from the National Cancer Institute's Surveillance, Epidemiology and End Results (SEER) (24) program were used in addition to NHIS data to generate expected numbers of events for comparison with the TCE Subregistry observed numbers. The SEER data can be used to generate rates that are strictly a measure of incidence (new cases in a particular year) and covers all reported cancers. To make the TCE Subregistry database comparable and to generate similar rates for comparison, the year of first treatment was used as the date of onset or the year of "new case."
Both the Registry and the SEER program are complete enumerations or data on an entire population. In the first case, the TCE Subregistry is the population of citizens exposed (and documented) to TCE in the environment; while in the latter case, SEER is the population of all (invasive) cancer cases in the 10 geographic areas constituting the SEER study area. Neither involves probability based sampling. Thus, comparisons that were made are descriptive rather than inferential in nature, and lend themselves directly to deterministic decision making.
The sparsity of the data for specific cancers in each age category, particularly the younger age groups, precludes statistical comparisons for specific cancers or even for specific age groups. The rates for total cancers by sex and year reported were calculated. No statistical comparisons were made; the observed and expected values are provided.
RESULTS OF DESCRIPTIVE COMPARISONS
This section provides a discussion of the comparability of descriptive data for the TCE Subregistry file and the NHIS file. Also, the results of this section were used to plan the subsequent analyses of the health outcome data. That is, the results were used to determine what variables were appropriate to include as covariates in modeling the health outcome comparisons.
Demographics
Race
In the NHIS sample, 18% (n = 21,066) of the subjects were nonwhite. In the TCE Subregistry sample, 3% (n = 127) of the subjects were nonwhite or missing (at next update, one missing was changed to white). Given this discrepancy in the proportion of nonwhites and the diversity of races reported among the nonwhites in the TCE Subregistry, all nonwhite subjects from the NHIS, NHSDA, and TCE Subregistry data were excluded from the analysis reported in this and subsequent sections of the report. The statistical analyses included 3,914 exposed, white TCE registrants.
Age
Table 5-3 shows the percentage of people in each age group. As expected, the percentages are affected by the timing of data collection relative to the exposure period. For example, the percentage of people in the 9 years of age or younger group is less for the TCE Subregistry than it is for the NHIS sample. For some of the TCE sites, the exposure period ended several years prior to data collection and precluded the eligibility of young children (that
Table 5-3.--Comparison of NHIS-TCE Subregistry population age groups.
Age Group (Years) |
NHIS | TCE Subregistry | ||
Number* | % of Total |
Number* |
% of Total |
|
All | 92,987 | 100.0 | 3,914 | 100.0 |
4 | 6,860 | 7.4 | 129 | 3.3 |
5 - 9 | 6,764 | 7.3 | 260 | 6.6 |
10 - 17 | 10,383 | 11.2 | 522 | 13.3 |
18 - 24 | 8,746 | 9.4 | 453 | 11.6 |
25 - 34 | 15,377 | 16.5 | 842 | 21.5 |
35 - 44 | 14,280 | 15.4 | 640 | 16.4 |
45 - 54 | 9,858 | 10.6 | 389 | 9.9 |
55 - 64 | 8,476 | 9.1 | 318 | 8.1 |
65 - 74 | 7,397 | 8.0 | 231 | 5.9 |
75 | 4,846 | 5.2 | 130 | 3.3 |
* Included only those reporting race as white.
is, children born after the exposure period ended). For some sites, exposure was occurring at the time of baseline data collection, and ensured the potential eligibility of children who were 9 years of age or younger at the time of the interview.
Table 5-3 also indicates, as expected, that there were fewer TCE Subregistry members in the older age groups. This pattern was consistent with ATSDR knowledge about the demographics of these TCE sites. For example, at one site the site address houses were built at about the same time as exposure began. These homes were inhabited primarily by young married couples. Therefore, the residents of this site were younger, on the average, than the general population.
The mean age for the NHIS sample was slightly greater than the mean of the TCE Subregistry sample (NHIS mean = 35.4 years, standard deviation (sd) = 22.3 years versus TCE mean = 34.0 years, sd = 19.9 years). This slight difference could reflect the conditions described in the previous paragraph.
Table 5-4 provides a comparison of the age-specific sex distribution of males in the TCE Subregistry and NHIS samples (the proportion of females can be calculated as one minus the proportion of males) and the corresponding p-values and confidence intervals for the differences between the TCE and NHIS proportions. Table 5-4 indicates that the proportion of males (like-
Table 5-4.--Age-specific distribution of white males.
TCE versus NHIS
Age | TCE* | NHIS* | p-value† | 95% CI§ | 99% CI§ |
0-4 | 0.48 | 0.51 | p = 0.56 | (-0.11, 0.06) | (-0.14, 0.09) |
5-9 | 0.53 | 0.52 | p = 0.54 | (-0.04, 0.08) | (-0.06, 0.10) |
10-17 | 0.49 | 0.51 | p = 0.28 | (-0.07, 0.02) | (-0.08, 0.03) |
18-24 | 0.48 | 0.49 | p = 0.71 | (-0.06, 0.04) | (-0.07, 0.05) |
25-34 | 0.47 | 0.50 | p = 0.08 | (-0.06, 0.01) | (-0.08, 0.01) |
35-44 | 0.48 | 0.49 | p = 0.96 | (-0.04, 0.04) | (-0.05, 0.05) |
45-54 | 0.51 | 0.49 | p = 0.52 | (-0.03, 0.07) | (-0.05, 0.08) |
55-64 | 0.46 | 0.48 | p = 0.57 | (-0.07, 0.04) | (-0.09, 0.06) |
65-74 | 0.40 | 0.45 | p = 0.13 | (-0.11, 0.01) | (-0.13, 0.03) |
75 | 0.43 | 0.37 | p = 0.16 | (-0.02, 0.15) | (-0.05, 0.17) |
*Cell values are proportions of total males for a given age group.
†The p-values are two-sided.
§The confidence interval is for the quantity: TCE proportion - NHIS proportion.
wise for females) did not differ significantly between the TCE and NHIS samples for any of the age categories examined. A marginally statistically significant (p = .08) difference was found for study members who were in the 25 through 34 years of age group. The findings in Table 5-4 suggest that sex ratios were similar for each age category in the comparison of TCE and NHIS populations.
Education Level
The summary statistics for education attainment are provided in Table 5-5. Because of the large number of comparisons involved with the assessment of age- and sex-specific differences between the TCE and NHIS samples for education, individual observed and expected counts are not provided. For both males and females, the overall difference in the age-specific education level of persons aged 19 years or older is statistically significant at the p 0.05 level. These overall differences were attributable to a larger proportion of TCE Subregistry members in the 9 though 11 years of education category (especially among younger adults), and a lower proportion who had at least a college degree compared with the NHIS sample. These findings indicate that the TCE Subregistry participants had fewer years of education; this difference could modify the comparison of health outcome rates.
Cigarette Smoking
Table 5-6 provides comparative information on current and past cigarette use for persons 18 years of age and older. Smoking rates are indicated for TCE Subregistry members and respondents to the NHSDA (15). Data are provided for all NHSDA respondents and for respondents in the North Central United States(1).
As with the NHSDA rates, TCE Subregistry rates (current smokers and ever smoked)(2) were higher for males than for females among persons 26 years of age and older. Among persons aged 18 though 25 years, in the TCE Subregistry slightly more females than males (55.9% versus 50.2%) reported ever having smoked. This pattern differed from the NHSDA rates, but was consistent with findings reported elsewhere (25).
Concerning the rate of current smokers, TCE Subregistry members had higher smoking rates in all age categories when compared with all NHSDA subjects. TCE Subregistry rates were more consistent with North Central United States NHSDA rates, with the greatest discrepancy among the 35 years of age and older group (34.6% versus 23.3%). The rates for all NHSDA age categories tended to be lower than the rates for the NHSDA North Central region cases, with the exception of the 35 years of age and older category. Sixty-one percent of all TCE Subregistry members who were 18 years of age and older were current or ex-smokers. This percentage was lower (61% versus 75%) than the percentage of "ever" smokers among all NHSDA respondents. The rates for ever smoked were substantially less among TCE Subregistry members for each age group: 18 through 25 years, 53.2% (TCE) versus 74.9% (NHSDA); 26 through 34 years, 63.1% (TCE) versus 80.8% (NHSDA); and 35 years and older, 62.4% (TCE) versus 79.2% (NHSDA).
Table 5-7 indicates the rate of ever smoked by education level for TCE Subregistry members and NHSDA respondents. The predominant pattern was a decrease in smoking with increased education for both males and females in all age groups. The rates for people with no
Table 5-5.--Mean years of education of NHIS and TCE Subregistry participants by age and sex.
(White Only, 19 Years of Age)
Condition | TCE Subregistry | NHIS | ||
Male | Female | Male | Female | |
Hypertension* | 7.0† | 10.2 | 9.9 | 12.5 |
Stroke§ | 1.3 | 2.0 | 1.0 | 1.2 |
Diabetes* | 2.7 | 3.3 | 2.3 | 2.8 |
Kidney disease* | 0.4 | 2.2 | 1.0 | 1.7 |
Urinary tract
disorders* |
1.7 | 5.5 | 1.5 | 1.3 |
Skin rashes* | 7.2 | 9.0 | 5.9 | 6.8 |
Anemia and other
blood disorders* |
1.8 | 4.0 | 0.6 | 2.3 |
Asthma, emphysema* | 4.8 | 6.4 | 8.6 | 10.8 |
Respiratory allergies* | 4.8 | 6.7 | 10.8 | 9.2 |
Stomach problems* | 5.3 | 9.0 | 6.2 | 9.2 |
Liver problems* | 0.3 | 0.4 | 0.3 | 0.2 |
Arthritis* | 4.6 | 9.3 | 13.8 | 19.1 |
Mental retardation¶ | 0.6 | 0.2 | 0.6 | 0.4 |
Speech impairment¶ | 1.2 | 0.6 | 1.0 | 0.6 |
Hearing impairment¶ | 2.6 | 2.0 | 10.9 | 7.5 |
*Indicates time frame is last 12 months.
†Percent of total population (white only).
§Indicates time frame is ever.
¶Indicates time frame is now have.
Table 5-9.--Summary of TCE Subregistry response rates for cancer.
Cancer | Sex |
Total | ||||
Male |
Female |
|||||
N |
% |
N |
% | N | % | |
None | 1,830 | 97.6 | 1,911 | 93.7 | 3,741 | 95.6 |
Lip, oral, pharynx | 2 | 0.1 | 1 | 0.0 | 3 | 0.1 |
Digestive system | 8 | 0.4 | 5 | 0.2 | 13 | 0.3 |
Respiratory system | 3 | 0.2 | 2 | 0.1 | 5 | 0.1 |
Unspecified site | 17 | 0.9 | 20 | 1.0 | 37 | 0.9 |
Breast | 0 | 0.0 | 23 | 1.1 | 23 | 0.6 |
Genital organs | 6 | 0.3 | 63 | 3.1 | 69 | 1.8 |
Urinary organs | 3 | 0.2 | 0 | 0.0 | 3 | 0.1 |
Lymphatic tissues | 4 | 0.2 | 1 | 0.0 | 5 | 0.1 |
Leukemia | 1 | 0.1 | 1 | 0.0 | 2 | 0.1 |
Other | 1 | 0.1 | 13 | 0.6 | 14 | 0.4 |
Total | 1,875 | 100.0 | 2,040 | 100.0 | 3,914 | 100.0 |
Controlling for Age and Sex
Table 5-10 provides a summary of the results of the NHIS and TCE Subregistry file comparison using Poisson regression analysis. For each health outcome, the table indicates the likelihood ratio statistics with the associated degrees of freedom and p-values for the effects of age (categorized into eight levels) and sex, based on a model containing age and sex. The residual deviance and the associated degrees of freedom are also given as a global lack of fit measure for this model, which specifies multiplicative effects of the age (i) and sex (j) ratios Oij/Eij. For each outcome, the age- and sex-specific numerators Oij are obtained from the TCE Subregistry data, while the expected numerators Eij are based on the suitably person-weighted age- and sex-specific ratios from the NHIS data. For the purpose of detecting structure in these age- and sex-specific ratios, a significance level of 0.05 is adopted.
As is shown in Table 5-10, the model was adequate but neither age nor sex was a statistically significant predictor in the models for skin rashes, stomach problems, and asthma and emphysema (p > 0.10). For both arthritis and respiratory allergies, the effects of sex were statistically significant while the effects of age were not. No test statistics are given for mental
Table 5-10.Summary of Poisson regression modeling.
Condition |
Age/Sex | Sex/Age | Residual Deviance
(p-value) |
df | ||
LR Stat*
(p-value) |
df† | LR Stat
(p-value) |
df | |||
Skin rashes | 10.93 (p = 0.14) | 7 | 0.58 (p = 0.44) | 1 | 7.96 (p = 0.33) | 7 |
Arthritis | 6.49 (p = 0.48) | 7 | 10.81 (p < 0.01) | 1 | 7.27 (p = 0.40) | 7 |
Mental retardation§ | --- | --- | --- | |||
Speech impairment¶ | 19.70 (p < 0.01) | 7 | 0.07 (p = 0.77) | 1 | 4.69 (p = 0.69) | 7 |
Hearing impairment | 75.95 (p < 0.01) | 7 | 0.34 (p = 0.55) | 1 | 4.83 (p = 0.67) | 7 |
Liver problems¶ | 23.34 (p < 0.01) | 7 | 3.28 (p = 0.06) | 1 | 6.65 (p = 0.46) | 7 |
Stomach problems | 11.74 (p = 0.10) | 7 | 0.82 (p = 0.36) | 1 | 11.32 (p = 0.12) | 7 |
Anemia and blood
disorders |
17.14 (p = 0.01) | 7 | 11.72 (p < 0.01) | 1 | 20.73 (p < 0.01) | 7 |
Diabetes¶ | 22.51 (p < 0.01) | 7 | 0.21 (p = 0.64) | 1 | 13.91 (p = 0.05) | 7 |
Kidney disease | 15.04 (p = 0.03) | 7 | 15.55 (p = 0.01) | 1 | 13.03 (p = 0.07) | 7 |
Urinary tract disorders¶ | 33.28 (p < 0.01) | 7 | 24.23 (p < 0.01) | 1 | 12.88 (p = 0.07) | 7 |
Hypertension | 14.41 (p < 0.04) | 7 | 2.29 (p = 0.13) | 1 | 16.23 (p = 0.02) | 7 |
Stroke¶ | 14.26 (p = 0.04) | 7 | 0.25 (p = 0.61) | 1 | 7.62 (p = 0.36) | 7 |
Respiratory allergies | 10.01 (p = 0.18) | 7 | 7.39 (p < 0.01) | 1 | 10.49 (p = 0.16) | 7 |
Asthma, emphysema | 9.70 (p = 0.20) | 7 | 0.25 (p = 0.61) | 1 | 12.08 (p = 0.09) | 7 |
*LR Stat - Likelihood Ratio Statistic.
†df - degrees of freedom.
§Model does not converge.
¶Some infinite estimates obtained.
retardation because the data were too sparse for the model to converge. For all other outcomes considered (speech impairment, hearing impairment, liver problems, anemia, diabetes, kidney disease, urinary tract disorders, hypertension, and stroke), statistically significant variations in the ratios were seen as a function of age. In addition, statistically significant sex effects were present for anemia, kidney disease, and urinary tract disorders, and a marginally significant sex effects were seen for liver problems. The multiplicative model for anemia exhibited significant lack of fit. Marginally statistically significant lack of fit was seen for diabetes, kidney disease, urinary tract disorders, chronic respiratory problems, and hypertension, suggesting that the effects of age depended on sex for these outcomes.
Table 5-11 summarizes the observed and expected numerators of the prevalence rates and their ratios for each of the health conditions considered separately. The results for age- and sex-specific analyses are detailed in Appendices D-1 through D-17. For the descriptive analysis, a ratio of 0/0 was considered to be undefined. For models in which the sex effect was not statistically significant, summary estimates based on the age-specific ratios for the combined sexes were given. For outcomes in which, at most, one effect was statistically significant, summary estimated ratios based on the most parsimonious model that did not exhibit lack of fit were also given together with 99% Wald confidence intervals (CI). Reliable estimates of the standard errors of estimated risk ratios of zero could not be obtained, so no confidence intervals are presented whenever this occurred. For outcomes for which a parsimonious model was not adequate, 99% Cis are presented for the individual age- and sex-specific ratios. For the purpose of identifying statistically significant excesses, a significance level of 0.01 was used. A discussion of the significant results follows.
Skin Rashes, Eczema, or Other Skin Allergies
For skin rashes (see Appendix D-1 for details) most of the age- and sex-specific ratios were greater than one, meaning that the corresponding reported numbers of positive responses in the TCE Subregistry data exceeded those expected based on NHIS rates. A similar pattern was observed in the age-specific ratios. Because neither age nor sex was a statistically significant predictor, an adequate summary measure for rashes was the overall ratio of O/E = 1.28, which was statistically significantly elevated with a 99% confidence interval (1.11, 1.48).
Arthritis, Rheumatism, or Other Joint Disorders
The age- and sex-specific estimates for arthritis are shown in Appendix D-2. Most of the age- and sex-specific ratios were less than one, indicating that the reported prevalence of arthritis was generally lower in the TCE Subregistry data than expected based on the NHIS prevalence rates. The estimated sex-specific ratios were lower for males than for females (0.37 and 0.56, respectively) and both were statistically significantly lower than the nominal ratio of 1.
Mental Retardation
No summary estimates are given for mental retardation (see Appendix D-3 for details) because only 1 of the 14 reported cases in the TCE Subregistry data occurred in individuals 35 years of age and older and the model would not converge. Overall, there were fewer cases in the TCE Subregistry data than expected based on NHIS prevalence rates, although there was a two- to fivefold excess in several age categories (based on small numbers).
Table 5-11.Summary of observed and expected health outcomes using multivariate models.
Condition | Observed | Expected | Risk Ratio | 99% CI* |
Skin rashes | 318 | 247.70 | 1.28 | 1.11, 1.48 |
Arthritis | 277 | 571.70 | 0.48 | 0.42, 0.57 |
Mental retardation | 14 | 22.58 | 0.62 | 0.31, 1.23 |
Speech impairment | 36 | 29.49 | 1.22 | 0.79, 1.88 |
Hearing impairment | 89 | 316.30 | 0.28 | 0.21, 0.37 |
Stomach problems | 284 | 288.30 | 0.98 | 0.84, 1.14 |
Anemia and other
blood disorders |
114 | 58.28 | 1.96 | 1.53, 2.49 |
Diabetes | 118 | 87.40 | 1.35 | 1.06, 1.71 |
Kidney disease | 53 | 54.04 | 0.98 | 0.69, 1.40 |
Urinary tract
disorders |
145 | 47.64 | 3.04 | 2.46, 3.78 |
Hypertension | 341 | 381.83 | 0.89 | 0.77, 1.02 |
Stroke | 65 | 34.00 | 1.91 | 1.39, 2.63 |
Respiratory allergies | 227 | 381.30 | 0.60 | 0.52, 0.71 |
Asthma, emphysema | 219 | 375.90 | 0.58 | 0.49, 0.69 |
Cancer | 55 | 45.59 | 1.21 | 0.85, 1.71 |
*CI - Confidence Interval for risk ratio.
Speech Impairment
For speech impairment, there was statistically significant variability in the age-specific observed to expected (O/E) ratios, with excesses in the TCE Subregistry data for people under 35 years of age, and deficits for people aged 35 years and older (see Appendix D-4 for details). Only the estimated risk ratio of 2.45 in the age group 0 though 9 years was statistically significantly elevated at the p 0.01 level. No confidence interval is given for the age-specific estimate for the 55 through 64 years of age group because the observed numerator of zero corresponded to an estimated parameter of negative infinity with an unreliable standard error.
There was a concern that the results might be directly related to differences in regional services provided when a child begins school, particularly given the smaller percentage of registrants in the 0 through 4 years of age group in the TCE Subregistry group than in the comparable NHIS group. Therefore, the analysis was repeated excluding children younger than 5 years of age. The results were similar for the 5 through 9 years of age groups (risk ratio = 2.41, 99% CI = 1.21, 4.80).
Hearing Impairment
The age-specific risk ratios for hearing impairment are given in Appendix D-5. Excesses were observed in the TCE Subregistry data for the youngest two age groups (individuals under 18 years of age), and deficits for the other age groups relative to the NHIS population. The estimated risk ratio for the youngest age group (0 through 9 years of age) was significantly elevated at the 0.01 level, and the estimated risk ratios for people aged 25 years and older were all statistically significantly less than one.
As previously discussed in the section on speech impairment, there was a concern that the results might be directly related to differences in regional services provided when a child begins school, particularly given the smaller percentage of registrants in the 0 through 4 years of age group in the TCE sample than in the comparable NHIS group. Therefore, the analyses were repeated restricting the population to those 5 years of age and older. The results were not significant at the p 0.01 level (risk ratio = 1.78, 99% CI, 0.84, 3.75 compared with 2.13, 99% CI of 1.12, 4.06 when children less than 5 years of age are included). It is not known whether this reflects an actual difference or an adjustment of sample sizes, but the information should be considered in data interpretation.
Liver Problems
The age- and sex-specific ratios for liver problems are given in Appendix D-6. Overall, there was a slight excess of reported cases in the TCE Subregistry data. The ratio of 11.63, based on two cases among females aged 55 through 64 years, was statistically significantly elevated at the p 0.01 level.
Ulcers, Gall Bladder Trouble, and Stomach or Intestinal Problems
The results for ulcers (the details appear in Appendix D-7) showed that the effects of age and sex were not statistically significant for this outcome. The marginal significance of the age effect was due to the apparent excesses in the youngest three age groups (under 25 years of age) and apparent deficits thereafter. The estimated overall risk ratio of 0.98 indicates an overall deficit in reported cases in the TCE Subregistry population compared with the cases in the NHIS population, although this deficit was not statistically significant.
Anemia or Other Blood Disorders
The age- and sex-specific risk ratios for anemia are detailed in Appendix D-8. No summary estimates are presented because the estimated effect of sex depends upon the age group. The largest excesses were seen in males. All ratios above 1.8 were statistically significantly elevated at the p 0.01 level. Ratios of this magnitude were seen in seven age-sex categories. The analysis was run using only those people 5 years of age and older; the results were similar. The risk ratio for males was 14.38 (99% CI = 3.97, 52.16) in the 5 through 9 years of age group; none were reported for females.
Diabetes
For diabetes (details appear in Appendix D-9), risk ratios and standard errors could not be obtained for the youngest two age groups, for which there were no reported cases in the TCE Subregistry data. Statistically significant risk ratios of 4.26 and 2.06 were seen for females aged 18 through 24 years and 45 through 54 years, respectively, and nonsignificant excesses were seen in most of the other age groups for both males and females. The marginal lack of fit of a multiplicative model indicated that heterogeneity within age groups did not follow a consistent pattern with sex.
Kidney Disease
The age- and sex-specific risk ratios for kidney disease are given in Appendix D-10. Both deficits and excesses were seen, with no overall elevation. The risk ratio of 4.57 for females aged 55 through 64 years was statistically significantly elevated.
Urinary Tract Disorders,
Including Prostate Trouble
The age and sex-specific risk ratios for urinary tract disorders are given in Appendix D-11. Fivefold and greater excesses were seen in several age groups, including a twelvefold excess among females in the youngest age group (0 through 9 years) based on 11 cases. Excesses were seen in all age groups except in males 55 years of age and older. All age- and sex-specific ratios for both males and females less than 35 years of age were statistically significantly elevated, as were the ratios for females in all other age groups except for those 45 through 54 years of age.
The analyses were repeated restricting the populations to those 5 years of age and older; the results were similar. The risk ratio for females ages 5 through 9 years was 8.26 (99% CI = 2.89, 23.65).
High Blood Pressure (Hypertension)
The estimated age- and sex-specific risk ratios for hypertension are given in Appendix D-12. Both deficits and excesses were seen for specific sex-, age-specific subgroups; there was no overall difference.
The Effects of Stroke
The age-specific estimates for stroke are shown in Appendix D-13. Parameters could not be estimated for the youngest three age groups due to sparse data. Consistent excesses were seen in the age-specific risk ratios for those aged 25 years and older, and these excesses were statistically significant for those aged 35 through 54 years and 65 years and older.
Other Respiratory Allergies or
Problems, Such as Hay Fever
All age- and sex-specific risk ratios for respiratory allergies (details are shown in Appendix D-14) were less than one, indicating a lower than expected reported prevalence of this condition in the TCE Subregistry data. The sex-specific estimated risk ratios for respiratory allergies were each statistically significantly (p 0.01).
Asthma, Emphysema, or Chronic Bronchitis
The age- and sex-specific risk ratios are given in Appendix D-15; the age- and sex-specific estimates showed no increased prevalence of asthma in the TCE Subregistry sample. The estimated overall risk ratio of 0.58 (99% CI = 0.49, 0.69) indicated a statistically significantly decreased risk in this group relative to the NHIS population.
Cancers
Appendix D-16 shows the results for the "all cancer" outcome using the NHIS all cancer period prevalence rates that included both first and second cancers. Results based on the NHIS rates using only the first cancer were essentially identical and are not shown. There were a number of age- and sex-specific observed-to-expected ratios greater than 1.0, and a similar pattern was observed in the age-specific ratios. Because neither age nor sex was a statistically significant predictor, an adequate summary measure for all cancers was the overall ratio O/E = 1.21, which was not statistically significantly elevated based on the 99% confidence interval.
Summary
All of the results are summarized in Table 5-12. This table presents a summary of the statistically significant (p 0.01) risk ratios observed in the Poisson regression analysis. Results were grouped according to the structure found in the corresponding Poisson regression model.
Controlling for Age, Sex, and Education
The Poisson regression analysis, which included an adjustment for education, is detailed in Table 5-13 for all outcomes except cancer. As before, all analyses included adjustments for age and sex. This analysis was restricted to individuals aged 19 years and older for whom
Table 5-12.--Summary of results*.
1. Grand Mean, No Structure with Age or Sex.
Structure | Condition | |
Rash | Asthma | |
Overall Summary | 1.28 | 0.58 |
2. Sex Effect Only
Sex | Condition | |
Arthritis | Respiratory Allergy | |
Males | 0.37 | 0.49 |
Females | 0.56 | 0.69 |
3. Age Effect Only
Age
(years) |
Condition | ||
Speech Impairment | Hearing Impairment | Stroke | |
0-9 | 2.45 | 2.13 | |
10-17 | |||
18-24 | |||
25-34 | 0.32 | ||
35-44 | 0.08 | 3.11 | |
45-54 | 0.03 | 4.58 | |
55-64 | 0.23 | ||
65 | 0.27 | 1.70 |
Table 5-12.--Continued.
4. Age and Sex Effect
(a) Males
Age
(years) |
Condition | ||||
Liver Problems | Anemia | Diabetes | Kidney Disease | Urinary Tract Disorders | |
0-9 | 3.54 | ||||
10-17 | |||||
18-24 | 6.12 | ||||
25-34 | 6.38 | ||||
35-44 | 6.29 | ||||
45-54 | |||||
55-64 | 8.56 | ||||
65 | 6.81 |
(b) Females
Age
(years) |
Conditions | ||||
Liver Problems | Anemia | Diabetes | Kidney Disease | Urinary Tract Disorders | |
0-9 | 12.79 | ||||
10-17 | |||||
18-24 | 2.19 | 4.26 | 6.38 | ||
25-34 | 5.56 | ||||
35-44 | 1.91 | 5.25 | |||
45-54 | 3.68 | 2.06 | |||
55-64 | 11.63 | 4.57 | 3.20 | ||
65 | 2.62 |
Table 5-13.Summary of Poisson regression modeling, including adjustments for education.*
Condition |
Effect | Likelihood Ratio Statistic (p-value) | df† | Residual Deviance | df† |
Skin rashes | Grand Mean
+ Age + Sex/Age + Education/Age, Sex |
3.68 (p = 0.60)
2.09 (p = 0.15) 3.08 (p = 0.21) |
5
1 2 |
22.26 (p = 0.72) | 27 |
Arthritis | Grand Mean
+ Age + Sex/Age + Education/Age, Sex + Age.Education/Main Effects Model |
4.55 (p = 0.47)
11.06 (p < 0.01) 3.39 (p = 0.18) 17.19 (p = 0.07) |
5
1 2 10 |
35.02 (p = 0.14) | 27 |
Hearing
impairment |
Grand Mean
+ Age + Sex/Age + Education/Age, Sex |
17.59 (p < 0.01)
0.27 (p = 0.60) 2.19 (p = 0.33) |
5
1 2 |
22.63 (p = 0.70) | 27 |
Stomach
problems |
Grand Mean
+ Age + Sex/Age + Education/Age, Sex + Age.Education/Main Effects Model |
6.44 (p = 0.26)
0.21 (p = 0.65) 1.13 (p = 0.57) 18.66 (p = 0.04) |
5
1 2 10 |
48.58 (p < 0.01) | 27 |
Diabetes | Grand Mean
+ Age + Sex/Age + Education/Age, Sex + Sex.Age/Main Effects Model |
5.89 (p = 0.32)
0.05 (p = 0.83) 0.27 (p = 0.87) 12.47 (p = 0.03) |
5
1 2 5 |
41.92 (p = 0.03) | 27 |
Kidney disease | Grand Mean
+ Age + Sex/Age + Education/Age, Sex + Sex.Age/Main Effects Model |
7.15 (p = 0.21)
16.64 (p < 0.01) 9.28 (p = 0.01) 10.46 (p = 0.06) |
5
1 2 5 |
33.56 (p = 0.18) | 27 |
Urinary tract
disorders |
Grand Mean
+ Age + Sex/Age + Education/Age, Sex |
25.17 (p < 0.01)
34.04 (p < 0.01) 19.21 (p < 0.01) |
5
1 2 |
33.40 (p = 0.18) | 27 |
Table 5-13.Continued.
Condition | Effect | Likelihood Ratio Statistic (p-value) | df | Residual
Deviance |
df |
Anemia and
other blood disorders |
Grand Mean
+ Age + Sex/Age + Education/Age, Sex + Sex.Age/Main Effects Model Grand Mean + Age + Sex/Age + Education/Age, Sex + Sex.Education/Main Effects Model |
16.62 (p<0.01)
0.52 (p = 0.47) 0.56 (p = 0.76) 38.82 (p < 0.01) 8.28 (p = 0.02) |
5
1 2 5 2 |
72.26 (p < 0.01) | 27 |
Stroke | Grand Mean
+ Age + Sex/Age + Education/Age, Sex |
15.20 (p < 0.01)
0.43 (p = 0.51) 2.23 (p = 0.33) |
5
1 2 |
24.78 (p = 0.59) | 27 |
Respiratory
allergies |
Grand Mean
+ Age + Sex/Age + Education/Age, Sex |
4.02 (p = 0.55)
12.44 (p < 0.01) 0.36 (p = 0.84) |
5
1 2 |
20.59 (p = 0.81) | 27 |
Asthma,
emphysema |
Grand Mean
+ Age + Sex/Age + Education/Age, Sex + Sex.Age/Main Effects Model |
7.02 (p = 0.22)
0.39 (p = 0.53) 1.44 (p = 0.49) 11.30 (p = 0.05) |
5
1 2 5 |
35.89 (p = 0.12) | 27 |
*Restricted to registrants 19 years of age or older.
†df - degrees of freedom
education status was not missing. Education attainment is categorized into three levels according to the highest level achieved at the time of the survey: (1) less than high school graduate, (2) high school graduate, and (3) post-high school education. The youngest age group includes ages 19 through 24 years, and the older ages were grouped as in the preceding analyses, giving a total of six age groups (19 through 24, 25 through 34, 35 through 44, 45 through 54, 55 through 64, 65 years and older) for this education-adjusted analysis.
A summary of the statistically significant effects assessed in a forward-stepwise manner is given in Appendices D-18 through D-23. Age is entered into the model first, followed by sex and education. The residual deviance is given for this main effect model. Each possible two-way interaction is assessed in the model containing all three main effects. Only interactions significant at p 0.10 were provided in the table. It is important to note that when interpreting the p-values for this analysis, splitting the data out by a third dimension (education) essentially reduces the sample size for each individual comparison, relative to the previous analysis. For this reason, the power is considerably less for any individual comparison. For some outcomes, it was not possible to obtain informative confidence intervals because the data were too sparse.
As shown in Table 5-13, education was a statistically significant main effect at the 0.01 level for both kidney disease and urinary tract disorders. In addition, education was a significant modifier of the age effect for ulcers, and a marginally significant modifier of the age effect for arthritis. Education was a significant modifier of the sex effect for anemia. Because the focus of this part of the analysis was to assess the extent of confounding or effect modification by educational status, subsequent analysis will concentrate on those outcomes for which either the main effect of education or interactions involving education were statistically significant. A discussion of significant results follows.
Arthritis, Rheumatism, or Other Joint Disorders
The results for arthritis are detailed in Appendix D-17. Education was a marginally significant modifier of the age effect for this outcome. Deficits were seen for almost all age-sex-education classifications. These deficits were statistically significant at the 0.01 level for males and females aged 65 years and older with less than a high school education, for female high school graduates aged 65 years and older, and for male high school graduates aged 25 through 34 years and 55 through 64 years.
Ulcers, Gall Bladder Trouble, and Stomach or Intestinal Problems
The results for ulcers are given in Appendix D-18. There was no significant sex effect for this outcome. Therefore, the age-education specific estimates are presented. The estimated risk ratio of 3.04 for adults under 25 years of age with more than a high school education was statistically significantly elevated.
Anemia
The results for anemia (see Appendix D-19 for details) show statistically significant excesses of anemia for males 55 years of age and older for both Education Levels 1 (no high school diploma) and 3 (post-high school education). Statistically significant excesses were also seen for females aged 45 through 54 years without a high school diploma, and for female high school graduates aged 35 through 44 years.
Kidney Disease
The risk ratios for kidney disease are presented in Appendix D-21. A statistically significant excess of kidney disease (based on three events) was seen for females 55 through 64 years of age with a post-high school education.
Urinary Tract Disorders
The results for urinary tract disorders are given in Appendix D-22. Consistent excesses were seen for females in all age and education categories, except for females aged 65 years and older without a high school diploma. These excesses were statistically significant for women less than 25 years of age without a high school diploma, for all female high school graduates except those aged 45 through 54 years, and for all women with post-high school education except those 45 through 64 years of age. No statistically significant excesses were observed for males.
Summary
Although there were some education effects in these data, it did not appear that education confounded the analysis in any substantial way. Including an adjustment for education yielded consistent results with the age- and sex-adjusted analysis in that the deficits for arthritis persisted, as did the significant excesses seen for anemia, kidney disease, and urinary tract disorders. Only one significant excess (for ulcers) was identified in the education-adjusted analysis that did not appear in the unadjusted analysis. The reduced power of this education-adjusted analysis could have accounted for the previously identified significant excesses in some age groups that were no longer statistically significant when adjusted for education.
Comparison with SEER Data
Tables 5-14 through 5-16 show the calculated rates for the TCE Subregistry population and the SEER rates (whites only) by sex for the years 1973 through 1988 (the years for which SEER data were available) along with the relative risks. (Appendix D-17 shows cancers by age and sex.) As can be seen, for males the rates in the TCE Subregistry were smaller for some years and greater for other years than those recorded by SEER; there was no consistent pattern. For females, the TCE Subregistry rates considerably exceeded the rates of SEER for the years 1978 through 1988, except for the years 1985 and 1986, when the rates were almost equal. After 1978, for the total population, the relative risk was close to one or considerably exceeded one (1978 through 1982 and 1988).
An adjustment of the rates for survival until 1989 was made for the SEER data (a built-in adjustment for the TCE Subregistry data); therefore, there was no known bias when the SEER rates were compared to the TCE Subregistry rates. The increasing relative risk from 1979 through 1988 might reflect actual increases in cancer rates, perhaps related to beginning environmental exposure and latency periods, or the increasing values might reflect an increasing ability by the registrants to recall more recent events, or a combination of both. The small number of events in the early years makes recall for the subregistry data very problematic; the SEER numbers were recorded events as new cancers were diagnosed.
Table 5-14.--Comparison of TCE Subregistry and SEER cancer rates (males).
Year of Diagnosis | Count* | TCE Subregistry Rate† | SEER Rate§ | Relative Risk¶ |
1973 | 0 | 0.0 | 40.9 | 0.00 |
1974 | 1 | 52.8 | 55.3 | 0.99 |
1975 | 4 | 211.2 | 62.2 | 3.40 |
1976 | 1 | 52.8 | 69.1 | 0.76 |
1977 | 0 | 0.0 | 77.1 | 0.00 |
1978 | 1 | 52.8 | 82.3 | 0.64 |
1979 | 3 | 158.4 | 92.0 | 1.72 |
1980 | 1 | 52.8 | 101.9 | 0.52 |
1981 | 3 | 158.4 | 115.5 | 1.37 |
1982 | 4 | 211.2 | 124.9 | 1.69 |
1983 | 2 | 105.6 | 146.2 | 0.72 |
1984 | 3 | 158.4 | 161.3 | 0.98 |
1985 | 4 | 211.2 | 189.5 | 1.11 |
1986 | 5 | 264.0 | 217.1 | 1.22 |
1987 | 3 | 158.4 | 267.4 | 0.59 |
1988 | 6 | 316.8 | 320.5 | 0.99 |
*Number of male registrants reporting any cancers.
†Rate per 100,000 (not adjusted for age).
§Surveillance Epidemiology and End Results (SEER) rates (per 100,000) not adjusted
for age; adjusted for survival until 1989.
¶Calculated relative risks using the crude TCE Subregisty rate divided by the crude
Table 5-15.Comparison of TCE Subregistry and SEER cancer rates (females).
Year of
Diagnosis |
Count* | TCE Subregistry
Rate† |
SEER
Rate§ |
Relative
Risk¶ |
1973 | 1 | 50.8 | 105.4 | 0.48 |
1974 | 0 | 0.0 | 142.2 | 0.00 |
1975 | 4 | 203.4 | 162.8 | 1.25 |
1976 | 2 | 101.7 | 168.1 | 0.61 |
1977 | 2 | 101.7 | 167.5 | 0.61 |
1978 | 6 | 305.0 | 171.7 | 1.78 |
1979 | 5 | 254.2 | 176.5 | 1.44 |
1980 | 7 | 355.9 | 185.3 | 1.92 |
1981 | 6 | 305.0 | 200.6 | 1.52 |
1982 | 10 | 508.4 | 212.0 | 2.40 |
1983 | 6 | 305.0 | 230.5 | 1.32 |
1984 | 7 | 355.9 | 257.3 | 1.38 |
1985 | 6 | 305.0 | 288.3 | 1.06 |
1986 | 7 | 355.9 | 314.0 | 1.13 |
1987 | 11 | 559.2 | 357.8 | 1.56 |
1988 | 15 | 762.6 | 403.3 | 1.89 |
*Number of male registrants reporting any cancers.
†Rate per 100,000 (not adjusted for age).
§Surveillance Epidemiology and End Results (SEER) rates (per 100,000) not adjusted
for age; adjusted for survival until 1989.
¶Calculated relative risks using the crude TCE Subregisty rate divided by the crude
Table 5-16.--Comparison of TCE Subregistry and SEER cancer rates (all registrants).
Year of
Diagnosis |
Count* | TCE Subregistry
Rate† |
SEER
Rate§ |
Relative
Risk¶ |
1973 | 1 | 25.9 | 73.75 | 0.35 |
1974 | 1 | 25.9 | 98.6 | 0.26 |
1975 | 8 | 207.2 | 113.4 | 1.83 |
1976 | 3 | 77.7 | 119.5 | 0.65 |
1977 | 2 | 51.8 | 123.2 | 0.42 |
1978 | 7 | 181.3 | 127.9 | 1.42 |
1979 | 8 | 207.2 | 135.1 | 1.53 |
1980 | 8 | 207.2 | 144.5 | 1.43 |
1981 | 9 | 233.1 | 158.9 | 1.47 |
1982 | 14 | 362.6 | 169.3 | 2.14 |
1983 | 8 | 207.2 | 189.2 | 1.10 |
1984 | 10 | 259.0 | 210.2 | 1.23 |
1985 | 10 | 259.0 | 239.8 | 1.08 |
1986 | 12 | 310.8 | 266.4 | 1.17 |
1987 | 14 | 362.6 | 313.4 | 1.16 |
1988 | 21 | 543.9 | 362.7 | 1.50 |
*Number of male registrants reporting any cancers.
†Rate per 100,000 (not adjusted for age).
§Surveillance Epidemiology and End Results (SEER) rates (per 100,000) not adjusted
for age; adjusted for survival until 1989.
¶Calculated relative risks using the crude TCE Subregisty rate divided by the crude
SEER rate.
1. 1 A descriptive summary of the TCE Subregistry members' current (at time of survey) and former use of cigarettes and other tobacco products is provided in Section 4 of this report.
2. 2 See page 43 of this report for definitions of the "current" and "ever" rates.