To determine whether pediatricians in managed care settings adhere to national guidelines concerning the provision of clinical preventive services.
Surveys were mailed between September 1996 and April 1997 to all pediatricians practicing in a California group-model health maintenance organization. The survey asked pediatricians about their screening and education practices on 34 recommended services and the actions taken with adolescent patients who have engaged in risk behavior.
The response rate was 66.2% (N = 366). Pediatricians, on average, screened 92% of their adolescent patients for immunization status and blood pressure; 85% for school performance; 60% to 80% for obesity, sexual intercourse, cigarette use, alcohol use, drug use, and seat belt and helmet use; 30% to 47% for access to handguns, suicide, eating disorders, depression, and driving after drinking alcohol; fewer than 20% for use of smokeless tobacco, sexual orientation, sexual and physical abuse, and riding a bike or swimming after drinking alcohol; and 26% to 41% for close friends' engagement in risk behavior. Pediatricians' assessment and education with adolescent patients who screened positive for risk behavior was particularly low. Female physicians, physicians who saw a greater proportion of older adolescents, and recent medical school graduates were more likely to provide preventive services.
Pediatricians in this health maintenance organization provide preventive services to adolescent patients at rates below recommendations but at rates greater than physicians in other practice settings. Improvement is especially needed in the areas that contribute most to adolescent mortality and for patients who screen positive for a risk behavior.
MOST ADOLESCENT morbidity and mortality can be attributed to preventable factors, such as alcohol, tobacco, and other drug use (ATOD); unsafe sexual practices; risky vehicle use; and weapons use.1 Although approximately 72% of adolescents visit a physician at least once a year, few are screened for or educated about these health risks. Consequently, national guidelines concerning physicians' provision of comprehensive preventive services to adolescent patients have been developed (eg, Guidelines for Adolescent Preventive Services2; Bright Futures: Guidelines for Health Supervision of Infants, Children and Adolescents3; Health Supervision Guidelines4; The Clinician's Handbook of Preventive Services: Put Prevention Into Practice5; and the Guide to Clinical Preventive Services6). In general, these guidelines recommend that all adolescents have an annual, confidential preventive services visit during which primary care physicians screen, educate, and counsel adolescent patients on a number of biomedical, emotional, and sociobehavioral areas currently threatening adolescent health.
Results of the few studies examining physicians' provision of a variety of preventive services to adolescent patients in private practice settings indicate rates far below recommended levels. For example, a review of data from the National Ambulatory Medical Care Survey found that only 1% of adolescent office visits included advice about smoking cessation or counseling about human immunodeficiency virus (HIV) transmission. Moreover, in about two thirds of adolescent visits, physicians did not counsel adolescents on weight and cholesterol level reduction, smoking cessation, HIV transmission, or breast self-examination.7 Marks and colleagues8 found that fewer than half of the pediatricians in private practice who were surveyed routinely provided anticipatory guidance about drug and alcohol use, sexuality and birth control, peer and family problems, depression, and incest and child abuse; 57% provided guidance regarding smoking; 80% discussed puberty and menstruation; and approximately 60% discussed school, career, and nutrition with adolescent patients. Studies9,10 that assessed provision of preventive services in greater depth within a specific service area (eg, sexual behavior or tobacco use) report similarly low rates of service provision.
Only one study9 has addressed the extent to which physicians provide follow-up services to adolescents who screen positive for risk behavior. This study focused only on screening of sexually active adolescents, finding that physicians in private practice educated approximately 70% of adolescents about sexually transmitted diseases (STD) and HIV transmission; asked 31% to 45% about sexual orientation, frequency of casual sex, and number of previous partners; and provided 10% with condoms.
Only one study11 has sampled physicians in a managed care setting. This study, which focused exclusively on preventive services related to sexual behavior, found that fewer than half of the sexually active female adolescent patients received an annual Papanicolaou smear and fewer than 10% were tested for an STD. No study, to our knowledge, has examined the extent to which physicians practicing in managed care settings provide the breadth of preventive services recommended for adolescent patients. This setting is of particular importance because enrollment in managed care during the past 15 years has increased dramatically, staff physicians can typically provide more preventive services at reduced or no cost to the patients, and most screening and diagnostic materials are now available on site.11
The present study assesses (1) the extent to which pediatricians in a group-model health maintenance organization (HMO) screen and educate adolescent patients concerning a number of health risks, (2) rates of follow-up screening and education to adolescents who report engaging in a risk behavior, (3) the extent to which pediatricians educate parents of adolescent patients, and (4) whether sources of variation in service provision in private practice settings (ie, patient age,10,12,13 physician sex,9,10 and years since medical school graduation9) similarly account for variation in service provision among managed care physicians. It is hypothesized that rates of screening and education will be below recommended levels and that female physicians, more recent graduates, and physicians seeing more older adolescent patients will report providing preventive services to adolescent patients at greater rates than their colleagues.
The study was initially designed to collect baseline data on pediatricians' practices to aid in the selection of sites to be used in the implementation and evaluation of a model for providing adolescent clinical preventive services in an HMO. All primary care pediatricians practicing in pediatric clinics within a group-model HMO in Northern and Southern California were recruited by mail. Physicians who did not specialize in pediatrics, who reported seeing fewer than 1 adolescent patient (aged 12-18 years) in the past 6 months, or who were pool or moonlighting physicians were excluded from the study.
All pediatricians practicing in this HMO were mailed packets in September 1996 (Northern California) and December 1996 (Southern California) that contained a cover letter explaining the primary purpose of the study, a letter endorsing the study from the chair of the Chiefs of Pediatrics, a consent form, an anonymous questionnaire, and an addressed and stamped return envelope. Approximately 2 months later, nonresponders were mailed a second questionnaire and a reminder letter. Another 2 months later, an additional reminder letter and questionnaire were mailed to the remaining nonresponders, and follow-up telephone calls were made within 1 week of this third mailing. Physicians were not compensated for their participation. All procedures were approved by the internal review boards at the University of California, San Francisco, and the participating HMO.
Of 809 packets mailed to physicians, 33 were undeliverable because the physician had moved, retired, or died; also, 73 physicians were excluded because they were pool or moonlighting physicians. Of 703 eligible physicians, 465 returned the questionnaire, for a final response rate of 66.2%. Of the returned questionnaires, 63 were excluded because the physicians reported not seeing any adolescent patients within the last 6 months, 4 were excluded because the physicians were not pediatricians, and another 32 were excluded because of excessive missing data, leaving a final sample of 366 physicians.
Physicians reported their sex; year of graduation from medical school; percentage of time in the clinic; number of adolescent patients seen per day and in the past month; percentage of adolescent patients seen in the past month aged 12 to 13, 14 to 15, 16 to 18, and older than 18 years; whether they encouraged an annual preventive service visit; and whether they provided confidential health care to their adolescent patients.
Physicians were asked about their provision of preventive services to adolescent patients during routine visits, defined as "non–acute care visits, such as routine checkups, sports examinations, and school or employment physicals." Physicians indicated the percentage of adolescent patients and their parents whom they screened and educated on 34 recommended services. Physicians also responded to a set of more detailed questions concerning their practices with patients who screened positive for tobacco use, alcohol use, sexual activity, and nonuse of a seat belt or helmet. Physicians responded to each question on a 0% to 100% scale divided into deciles (ie, 0%-10%-20%-30%, and so on). Similar questions and response scales have been used in other studies9 of physicians' provision of preventive services.
Means and SDs were calculated to describe physicians' reports of the average percentage of adolescent patients who received each of the recommended services and the percentage of physicians who provided each service to all adolescent patients. Bivariate correlations were computed to assess the relation between the proportion of younger adolescents seen (ie, aged 12-13 years) and the average percentage of adolescents screened and educated. To limit the number of analyses, this was conducted only on the practice areas accounting for the largest amount of adolescent morbidity and mortality. Hierarchical linear regression analysis was used to assess whether physician sex and years since medical school graduation accounted for variation in pediatricians' provision of an unweighted average of the recommended follow-up practices for adolescent patients who had already engaged in a risk behavior. Because the average number of years since medical school graduation was less for female than male physicians, it was necessary to control for one physician variable when examining the effects of the other. This analysis was performed by entering the control variable at the first step and the physician variable of interest at the second step. All statistical analyses were conducted using a software program (SPSS for Windows release 7.5; SPSS Inc, Chicago, Ill).
As indicated in Table 1, the final sample of 366 pediatricians comprised 193 men and 168 women; 5 physicians did not report their sex. Participants had, on average, completed medical school 20 years earlier. Pediatricians reported spending an average of 84% of their time working in the pediatric clinic and seeing an average of 76 adolescent patients in the past month. On average, 43% of the adolescent patients were aged 12 to 13 years, and only 3% of patients were older than 18 years. Twenty-four pediatricians (6.6%) reported having had specialty training in adolescent medicine. Thirty-two percent of physicians encouraged adolescent patients to have an annual preventive services visit. Almost all physicians (92%) provided confidential care to their adolescent patients, with 88% making their policies about confidentiality clear to adolescents and 71% making their policies clear to parents.
Table 2 presents the average percentage of adolescent patients whom physicians reported screening in each of the 24 recommended services. (Two separate analyses were conducted—1 with and 1 without the 24 pediatricians reporting adolescent medicine specialty training. Because both analyses yielded the same pattern of results [data not shown] and the focus of this article was on preventive services among all pediatricians practicing in HMOs, our data include these specialists.) On average, pediatricians reported screening the greatest percentage of adolescents for immunization status, blood pressure, and school performance and the fewest for use of smokeless tobacco, sexual orientation, sexual and physical abuse, and alcohol use while riding a bike or swimming. The pediatricians typically screened 60% to 80% of adolescents for obesity, ATOD, sexual intercourse, and consistency of seat belt and helmet use and 26% to 47% for driving under the influence of alcohol, depression, eating disorders, suicide, access to handguns, and close friends' and families' risk behavior. The SDs for all of these average percentages were large, indicating a broad range of service provision reported by pediatricians.
Published guidelines (eg, Guidelines for Adolescent Preventive Services2) suggest that physicians screen all of their adolescent patients on each of the areas listed in Table 2. Data from this study revealed that in none of the recommended areas of screening did every physician report screening all of their adolescent patients (Table 2). Because it might be unrealistic to expect physicians to consistently provide each service, physicians' usual screening practices, defined as providing the service to at least 80% of their adolescent patients, were examined. Results indicated that physicians usually provided a mean (SD) of 9.0 (5.4) of these 24 services, and 4% did not provide any of these services as part of their usual practice with adolescent patients.
The average percentage of adolescent patients whom physicians reported educating on each of the recommended service areas is listed in Table 3. On average, physicians reported educating 50% to 65% of adolescents about nutrition, STD and HIV transmission, birth control, risks of alcohol and drug use, injury prevention, and physical growth and fewer than half about psychosexual and psychosocial development or weapons. Physicians reported speaking with 78% of the parents of adolescent patients, discussing normative physical and psychosocial development with just more than 50% of these parents, and discussing monitoring of adolescent behavior and warning signs of emotional distress with fewer than half of the parents. As with screening, the large SDs in education practices reflect a broad range of service provision among sampled pediatricians.
Fewer than one third of the physicians in this study reported educating all of their adolescent patients or their parents on any of the recommended areas (Table 3). Of the 15 areas of education studied, physicians provided a mean (SD) of 5.0 (4.0) of these services to at least 80% of their adolescent patients.
Physicians' practices with patients who screened positive for risk behavior are summarized in Table 4. Physicians reported, on average, assessing 77% and 66% of adolescent patients who reported smoking about the amount smoked and family history of tobacco use, respectively, and educating 84% about the risks of tobacco use. Physicians typically asked just more than 70% of their adolescent patients reporting alcohol use about the amount and frequency of use, just more than 50% about family history of alcohol use or the incidence of driving after drinking alcohol, and fewer than 15% about their use of alcohol in conjunction with other activities. On average, physicians reported performing a pelvic examination and Papanicolaou smear on 33% of their sexually active female adolescent patients and screening 42% for an STD. Most of the other sexually active patients were referred to another clinic for these services. Physicians educated, on average, 70% of their sexually active adolescents about birth control methods, educated 80% about STD and HIV transmission, asked approximately half about the number of sexual partners and/or pregnancy history, provided 22% with condoms and 29% with some other birth control method, and assessed 29% about whether they felt pressured to have sex. On average, physicians reported educating almost 80% of their adolescent patients who reported inconsistent seat belt or helmet use about the associated risks. With a few exceptions, fewer than half of the physicians surveyed provided follow-up assessment and education to all of their adolescent patients who reported already engaging in a risk behavior.
There was a significant negative relation between the percentage of 12- to 13-year-old adolescent patients seen and rates of service provision. Physicians who saw a greater proportion of younger adolescent patients were less likely to ask about sexual intercourse, ATOD, access to handguns, consistency of seat belt use, and risk of suicide than were physicians who reported seeing a greater proportion of older adolescents (r values ranged from −0.12 [P<.05] to −0.22 [P<.01]).
Hierarchical linear regression revealed that after controlling for years since graduation, female physicians screened and educated tobacco-using (R2 change = 0.01; P = .03), alcohol-using (R2 change = 0.02; P = .01), and sexually active adolescent patients (R2 change = 0.05; P = .000) more often than did their male counterparts. Controlling for physicians' sex, more recent graduates screened and educated adolescent patients reporting tobacco use (R2 change = 0.02; P = .005), alcohol use (R2 change = 0.01; P = .06), and sexual activity (R2 change = 0.07; P = .000) more often than did less recent graduates. Combined, physician sex and years since graduation accounted for 5%, 4%, and 16% of the variance in tobacco-, alcohol-, and sexual behavior–related services, respectively (P<.001).
National guidelines2- 6,14 suggest that primary care physicians should play a critical role in preventing and reducing adolescents' participation in behaviors that place them at health risk by screening and educating all adolescent patients on a variety of biomedical, emotional, and sociobehavioral health risks. The present study is the first to examine rates of provision of a wide range of recommended preventive services to adolescent patients among pediatricians practicing in a group-model HMO.
Study results indicate that pediatricians in this HMO are establishing a positive milieu in which to provide preventive services. In accordance with national guidelines2 and a recent position paper from the Society for Adolescent Medicine,15 almost all of the physicians report providing confidential care to their adolescent patients, and they make these policies clear to the patients and their parents. Provision of confidential care to adolescent patients is critical if we want to increase adolescents' willingness to disclose important and often sensitive information to their physicians.16 In addition, despite the California-based HMO's guidelines that adolescents should have a preventive service visit every 2 to 3 years, about one third of the physicians in this study reported encouraging such visits, as recommended by national guidelines.2
Even with the established positive environment, results from this study indicate that rates of provision of preventive services to adolescent patients, as reported by the pediatricians practicing in this HMO, are still below recommended levels.2- 6 On average, pediatricians in this study reported screening most adolescent patients for biomedical services and school performance and about 70% for risk behaviors most associated with adolescent health problems: ATOD and sexual activity. Furthermore, pediatricians educate about half of their adolescent patients about STD and HIV transmission, birth control methods, risks of alcohol and substance use, injury prevention, and nutrition. Physicians report asking fewer than a quarter of their adolescent patients about sexual orientation, level of sexual activity among close friends, sexual or physical abuse, and use of smokeless tobacco.
Although variation in research methods makes comparisons across studies difficult, pediatricians in this HMO seem to be providing preventive services to their adolescent patients at higher rates than those practicing in other settings. For example, Marks et al8 found that 46% to 57% of the pediatricians in private practice studied provide anticipatory guidance regarding sexual activity and ATOD. Using a California sample, Ellen et al12 found that pediatricians screen, on average, 63% to 68% of their 11- to 14-year-old patients for sexual activity or ATOD. However, these rates increase to almost 80% for their 15- to 18-year-old patients. It is possible that the higher rates of screening reported by the HMO physicians in this study reflect greater emphasis on preventive services and access to on-site screening and diagnostic tools, many of which are provided at little or no cost to the patient.
There are 2 areas in which improvements in physicians' practices are particularly warranted. First, an increase in preventive service provision is needed in the areas that contribute most to adolescent mortality. National data1 indicate that unintentional injuries are the leading cause of adolescent death. However, physicians in this study reported screening, on average, fewer than 60% of adolescent patients for seat belt and helmet use, fewer than half for driving under the influence of alcohol, and 11% for alcohol use while riding a bike or swimming. National statistics1 indicate that homicide is the second leading cause of adolescent death, that 31% of boys and 8% of girls reported carrying some form of a weapon, and that 8% of adolescents reported carrying a gun within the past 30 days. Physicians in this study assessed access to handguns in fewer than one third of adolescent patients, and they educated, on average, 25% of patients about weapons. Finally, suicide is the third most common cause of adolescent mortality, with data1 indicating that 6% to 13% of adolescents attempt suicide at least once. Physicians in this study screened, on average, 46% and 35% of their adolescent patients for depression and suicide, respectively.
SECOND, IT IS concerning that pediatricians' average rates of provision of many of the recommended services to their adolescent patients already participating in risk behaviors is low. In particular, pediatricians do not provide condoms or other methods of birth control to many of their sexually active adolescents, and they do not assess whether adolescents ever feel pressured to have sex. Similarly, physicians assessed few patients who reported drinking alcohol about whether they drink while performing another activity, such as bike riding or swimming. These results indicate that screening does not necessarily lead to effective follow-up.
Consistent with results of studies9,10 of physicians in private practice, results of our study revealed that female pediatricians and more recent medical school graduates provide more preventive services than do their colleagues. Although the percentage accounted for in services provided to sexually active adolescents by these variables is small, the 7% accounted for in services provided to sexually active adolescents by years since graduation is substantially larger than reported by others (eg, 1% in the study by Millstein et al9). The significant relation between years since graduation and provision of preventive services likely reflects greater emphasis on prevention in more recent medical school curricula, and the relation between physician sex and prevention practices might reflect a greater comfort in discussing sensitive issues with adolescents, especially about sexuality, among female than male pediatricians.
Also consistent with past research,10,12 results of our study indicate that physicians who see a greater proportion of younger adolescent patients (ie, aged 12-13 years) than older patients report providing fewer preventive services. In contrast, a chart review13 revealed that 13- to 14-year-olds were screened for more biomedical and sociobehavioral risks than were 17- to 18-year-old patients. The findings from the present study might reflect a lack of awareness or a disbelief that adolescents at this young age are already engaging in risk behaviors. This lack of screening is unfortunate because evidence suggests that the average age of initiation of risk behaviors among adolescents is becoming increasingly younger.1
This study is limited by its use of self-report data. Past studies17 indicate that individuals exaggerate their behaviors in self-report questionnaires. In addition, given that the questionnaire was originally intended for site selection, physicians might have overestimated their reported practices because of their desire to participate in the intervention. However, the survey administered for this study was anonymous, thereby serving to reduce response bias and increase honesty. If physicians are overreporting their screening and educating practices, this would suggest that their true preventive service practices are even lower. Finally, this study did not address all of the obstacles believed to impede pediatricians' practices, including lack of comfort in talking with adolescents about sensitive topics such as sexual behavior,18 low self-confidence that prevention messages will be effective in changing adolescent behavior,10,19,20 or the small amount of time allotted for most routine visits (average, 15 minutes).1
In summary, our data reveal that although pediatricians in this HMO are providing many of the 34 recommended services to adolescent patients at greater rates than found in other practice settings, the rates are nevertheless below recommended levels.2- 6 Improvement in physicians' provision of preventive services is especially warranted in the more sensitive areas, in the areas most contributing to adolescent mortality, and for patients who have already screened positive for a risk behavior. The results strongly suggest that the recommendations might need to be prioritized because not all 34 areas are always possible or of equal importance within the typical office visit. Results also indicate that a systems change is needed whereby efforts are made to increase the amount of time allocated per patient visit, to increase the acceptance of additional resources for screening procedures even if additional costs are incurred, and to teach pediatricians the skills necessary to effectively screen and educate adolescent patients on a variety of key emotional and sociobehavioral areas.
Editor's Note: We still have a long way to go before we do well by adolescents. However, it's nice to see that the younger physicians are better at it than the older. There's hope.—Catherine D. DeAngelis, MD
Accepted for publication June 21, 1999.
This work was supported primarily by grant 96-42 from The California Wellness Foundation. Additional support was provided by grant 2T71MC0000321 from the Maternal and Child Health Bureau (MCHB), Health Resources and Services Administration (HRSA), Department of Health and Human Services, Rockville, Md, and by the National Adolescent Health Information Center, San Francisco, Calif, which is supported primarily by grant 1H06MC090000301 from the MCHB, HRSA, and the Department of Health and Human Services.
We thank Lauri Pasch, PhD, for her assistance in developing the physician survey; Robert Pantell, MD, and 2 anonymous reviewers for their insightful comments on the manuscript; Sally Ward, Scott Burg, and Holly Sigler for their skillful preparation of the manuscript; and the many pediatricians who took the time to complete this survey.
Reprints: Bonnie L. Halpern-Felsher, PhD, Division of Adolescent Medicine, Department of Pediatrics, University of California, San Francisco School of Medicine, 3333 California St, Suite 245, Box 0503, San Francisco, CA 94143-0503 (e-mail: email@example.com).
Halpern-Felsher BL, Ozer EM, Millstein SG, Wibbelsman CJ, Fuster CD, Elster AB, Irwin, CE. Preventive Services in a Health Maintenance OrganizationHow Well Do Pediatricians Screen and Educate Adolescent Patients?. Arch Pediatr Adolesc Med. 2000;154(2):173-179. doi:10.1001/archpedi.154.2.173