Screening children for blood cholesterol levels is currently recommended for those with positive family histories of premature cardiovascular disease.¹ Wider screening has not received much support. Recent literature suggests that childhood levels of cholesterol cannot be expected to decrease greatly with only dietary measures, and that untested longer adverse effects might ensue if lipid-lowering agents are used.^2,3 Opponents of wide screening of children for blood cholesterol have pointed out that concern over "elevated" levels might create unnecessary and unproductive anxiety in parents and children. Proponents of ascertaining childhood levels of cholesterol point to the relatively strong "tracking" of cholesterol over time and the failure of family history to detect elevated cholesterol levels in children or to elicit information on cholesterol levels in family members.⁴ The relationship of adult cardiovascular disease to cholesterol levels, as well as the presence of early atherosclerotic lesions in blood vessels of youth being correlated with childhood levels of low-density lipoprotein cholesterol,^7,8 have also been cited as rationale for knowing children's cholesterol levels.

The response of parents or physicians in practice to the recent National Cholesterol Education Program (NCEP)¹ recommendations is largely unknown. Previous surveys have indicated that a significant proportion of pediatricians were screening some children for cholesterol.^9,10 Until now, little information was available on parental actions and concerns and physician responses to parental notification that a child's cholesterol value was 200 mg/dL or greater, a value recommended by the NCEP to warrant physician follow-up and evaluation.¹

The occurrence of a large, four-region (San Diego, CA; New Orleans, LA; Minneapolis, MN; and Austin, TX) trial of a school-based intervention provided an opportunity to assess the impact on parents, children, and physicians of the knowledge that a child had a total blood cholesterol value found to be elevated (200 mg/dL or greater, according to NCEP guidelines). The total blood cholesterol level measurements for this study were obtained as part of participation in a health trial rather than as a result of a screening procedure in a physician's office regarding cholesterol level. This would somewhat mimic the conditions that might pertain to a larger screening situation, however, in that after detection of an elevated value, further confirmation and follow-up are suggested.

Study questions addressed by this research include reported concerns or worries subsequent to the notification of an elevated level and any actions taken by the family. Actions taken could include consulting a physician about the test result and/or initiating dietary or other health habit changes. We also assessed the effects of geography, finances for medical care, ethnicity, gender, presence or absence of family history of heart disease, and study group assignment on actions taken by the parents. In addition, we assessed physician responses to consultation for this issue.

METHODS

Overview of Main Child and Adolescent Trial for Cardiovascular Health Trial

All CATCH cohort children's parents were mailed a letter after each assessment, at baseline, and at follow-up, which gave the results of the child's height, weight, average systolic blood pressure, average diastolic blood pressure, and total blood cholesterol value. The following paragraph was included in every letter; the intent of the investigators was to emphasize the recommendation, but not to cause undue alarm:

Total blood cholesterol: 205 mg/dL. Current recommendations are that blood cholesterol be below 200 mg/dL. However, a single measurement is not adequate to make a diagnosis of high blood cholesterol. Thus, if your child's value is above 200, we recommend that you see your own physician for further testing and advice.

The letter was signed by the principal investigator at each study site, and a phone contact for the CATCH site investigator was given if a parent had questions. Lists of all children and their values were given to each CATCH site principal investigator for handy reference should a parent or physician call. Principal investigators reported few (generally less than 10) calls at each site. The school nurse, if present in a CATCH study school, was also given a list to follow up with the parents.

Measurement and Definitions

Between 3 months and 6 months after the final fifth-grade measurement of the cohort, a telephone survey was conducted by the University of Minnesota Phone Survey Unit. The period between notification of a parent and the telephone survey could range between 3 months and 3 years, depending on which measurement (third or fifth grade) was pertinent. Some parents had children with levels of 200 mg/dL or greater at both third and fifth grades and would have received two notices. Parents were initially contacted; their physicians were contacted after parental permission had been granted. Interviewers were trained and supervised by the Division of Epidemiology, University of Minnesota. The content for the 30- to 40-minute interview included: parental recall of and action taken regarding notification at either the third or fifth grade of a child's cholesterol level being 200 mg/dL or greater; demographic data; perceived levels of concern (parents and child); cholesterol testing of parents; recent changes in diet and exercise habits; and medical insurance information. Study group assignment was already in the database. Pretesting assured respondent feasibility and reliability of survey instruments. Interviews were completed using a computer-assisted telephone interviewing software package available from the University of California (Berkeley, CA). The primary reason for not completing an interview was the inability to locate and reach the family after no less than 3 but up to 10 attempts were made.

Statistical Methods

Multiple logistic regression, with adjustment for random effects of school, was carried out using an SAS version 6.10 macro procedure to assess the strength of these associations further and to explore which combined factors best predicted the outcome (GLIMMIX, a macro for fitting generalized linear mixed models, unpublished procedure, SAS Institute, Inc, Cary, NC).¹⁵ For the multiple logistic regression, the child's cholesterol level in two determinations (if that was the case) was averaged. Because of small sample sizes of some ethnic groups, the logistic regression was limited to children with white, African-American, and Hispanic ethnicities.

RESULTS

Parent Characteristics and Parental Knowledge and Attitudes Concerning Notification

Among the parent respondents, 538 (69%) were notified of their children's high cholesterol results once in either third or fifth grade, and 246 (31%) were notified twice. Among those notified once, 84% recalled receiving the notice, yet only 39% of those remembered the results as being high. Fifty-eight percent indicated that the results caused them at least some concern. In comparison, 88% of those receiving two notices recalled ever receiving any notification, and 77% of those indicated the results had caused them concern. Although more than one third (34% to 37%) of the respondents who had concerns about their children's cholesterol levels indicated they were "very or extremely concerned," less than 22% thought it caused the same level of concern in their partners or in the children themselves.

Characteristics of the Physicians' Responses

Table 1. Child and Adolescent Trial for Cardiovascular Health (CATCH) Ancillary Study: Physician Survey Response (Sample = 179) [View Table]

Although no physicians indicated that medication had been prescribed, two parents reported that medication (type not specified) had been recommended for their children. No question was asked regarding compliance or duration of treatment. One child (in the third grade) had a cholesterol level of 210 mg/dL and a follow-up level (at fifth grade) of 169 mg/dL; the other child had a (third grade) cholesterol level of 239 mg/dL and a follow-up level (fifth grade) of 215 mg/dL.

Characteristics Associated with Consulting a Physician

Considering these factors, characteristics of stronger interest and more plausible correlation with the outcome were selected for inclusion in the logistic regression model (Table 2). After adjusting for race and site, parents who were notified twice were twice as likely to consult physicians (P < .001). Families whose adult respondents to the survey had themselves been tested for cholesterol were close to twice as likely to consult physicians for their children than those who did not (P < .01). In addition, the higher the child's cholesterol level, the more likely parents were to seek physician consultation (P < .05). Respondents who had some college education tended more to consult physicians than those who had only high school education or less (P < .05). Those having health insurance were twice as likely to take the opportunity to see physicians. Family history, when other factors were present, was not a significant predictor (P = .8). Geographic location was also not significant (P = .07).

Table 2. Odds Ratio for Consulting a Physician After Notification of Child's High Cholesterol, From Multiple Logistic Regression [View Table]

DISCUSSION

The figure may be somewhat altered because of failure of recalled information. There also could be potential respondent bias to report consulting a physician. This is suggested in the finding that several physicians denied being consulted by the parents about the problem. Failure to recall information after a long period may also explain not only the substantial proportion of parents reporting not recalling the notification but also a considerable proportion not recalling the values as being high. It may be that parents had less concern because the reason for obtaining the blood test in the first place did not originate with their own physicians but, rather, originated as a side effect of participation in a research project. It also may be the case that parents do not associate blood cholesterol with increased risk of cardiovascular disease in this younger age group. They may not believe that much can be done medically at this time in response to a cholesterol value noted in children. Parents, however, may believe that lifestyle habits and behaviors such as not smoking, diet choices, and activity are more important than cholesterol levels in this age group. Luepker¹⁹ arrives at this conclusion based on his review of attempts to improve youth health habits. Not surprisingly, the data show that higher actual values and being notified twice were significantly associated with following the NCEP guideline recommendations. Parental concern about their own or their partners' cholesterol levels was also associated, whereas the presence of a family history of cardiovascular disease, in the presence of the other predictive factors, was not a significant factor.

Analysis of physician responses to presentation of a child with an elevated value of cholesterol also shows uncertainty about complete compliance with the NCEP guidelines. After seeing the children, only 60% of physicians reported evaluating them, and only half reported repeating the test, which is the first recommendation. Not all inquired about family history, although it is possible that this had already been ascertained. Nearly 40% made dietary recommendations, but only a few made referrals to dietitians. No physician contacted admitted recommending medication, but two parents thought that medication had been recommended by physicians we were unable to contact. This finding is cause for some concern, in that NCEP guidelines generally do not recommend medication at this first stage of investigation and level of cholesterol.

CONCLUSIONS

We speculate that the emphasis of messages (for the lay public) related to adult cardiovascular disease prevention with youth should be on establishing healthy dietary habits and other positive health behaviors, such as not smoking and enhanced physical activity, rather than focusing on the blood cholesterol level. These population-based recommendations have been advanced by the NCEP as well. The public health effects, such as those demonstrated in the main CATCH study of influencing the school environment to promote healthful behaviors in youth, deserve more emphasis and are likely to attract more professional and lay public support.