Direct Observation of Health-Habit Counseling of Adolescents FREE

THE HEALTH status of adolescents in the United States continues to decline as a result of the increasing prevalence of risky health behaviors, specifically, alcohol, drug, and tobacco use; poor diet; and unsafe sex.^1- 11 Engaging in these behaviors puts young people at risk for serious morbidity and mortality in both the short and the long term. In addition, many of the risk factors for morbidity and premature death in adulthood stem from risky health behaviors that begin during adolescence.

Because most adolescents (70%) visit a physician at least once per year,¹² physician visits provide an opportunity for these young people to receive preventive services that not only can improve their current state of health but also may provide the opportunity to prevent substantial morbidity and mortality later in life. Most outpatient visits by adolescents are to family physicians.¹³ The rates of preventive service delivery to adolescents in family practice have been found to be comparable to those seen in pediatric practices.¹⁴

To help physicians address the physical and mental health needs of their adolescent patients, the American Medical Association developed a series of prevention recommendations for adolescents aged 11 to 21 years, Guidelines for Adolescent Preventive Services (GAPS).¹ The publication of GAPS helped to eliminate one of the commonly cited barriers to preventive service delivery for these young people, specifically, the lack of clear, consistent recommendations for prevention in this age group.¹⁵ Similarly, preventive care needs of adolescents have been addressed in the clinical preventive service recommendations made by the US Preventive Service Task Force,¹⁶ the American Academy of Family Physicians,¹⁷ the American Academy of Pediatrics,¹⁸ and the Maternal and Child Health Bureau of the Health Resources and Services Administration.¹⁹ These recommending bodies are consistent in their emphasis on improving the prevalence and quality of preventive service delivery to adolescents, and, for the most part, are in agreement on which counseling services should be delivered.²⁰ Yet, the publication and dissemination of recommendations alone do not ensure that recommended services will be delivered. The surveillance of prevention in the primary care setting is needed to ascertain the actual rate of preventive service delivery.

Many studies have examined various aspects of preventive care delivery to this age group, including the role of community health centers²¹; the importance of confidentiality^22- 23; the cost of comprehensive preventive service delivery¹²; and the delivery of specific preventive services such as counseling on tobacco use,^24- 25 cholesterol screening,²⁶ and injury prevention counseling by nurse practitioners.²⁷ A paucity of information, however, still exists on many aspects related to preventive service delivery to adolescents.^28- 29

A few published studies^14- 15 regarding the rates of delivery of recommended preventive services to young people have found that the rates were far below those recommended by GAPS. The methods of data collection used in both studies, however—physician and patient self-report and medical records review—have been shown^30- 33 to seriously underreport health habit counseling. This is particularly important for assessing the delivery of recommended adolescent clinical preventive services that emphasize health-habit counseling.

Knowledge is also limited about the factors associated with preventive service delivery to this age group. A few studies have provided evidence that patient demographics have an effect on the delivery of preventive services, specifically, patient age,^14,24 race,¹⁵ and sex.¹⁴ Many of the findings, however, have been inconsistent. Furthermore, to date, no studies have examined visit or physician characteristics as they relate to preventive service delivery in adolescents.

The purpose of this study was, first, to report the rates of delivery of the GAPS-recommended preventive counseling services to adolescents during visits to family physicians, using the direct observation of patient visits to measure preventive health–habit counseling. Second, we wished to identify the characteristics of patients, office visits, and physicians that are associated with the delivery of health-habit counseling services. By identifying the factors associated with delivery, it may be possible to target policy changes, interventions, or both, that may increase the delivery of services to the adolescents who are currently not receiving appropriate counseling.

The methods of this study have been described previously in detail.^31,34 Briefly, a cross-sectional study of visits to family physicians in northeast Ohio was conducted from October 1, 1994, through August 31, 1995. Physician members of the Ohio Academy of Family Physicians practicing within 80 km of Cleveland or Youngstown were invited to participate in a study of the content of family practice. Of the 531 physicians who were invited to participate, 138 volunteered initially and were included in the study. Based on power calculations, a sample of 120 physicians was determined to be adequate to test the main hypotheses for the larger study. Physicians who participated were demographically similar to practicing members of the American Academy of Family Physicians on a number of characteristics, including physician age, percentage who were residency trained, and the mean number of patients seen per week. The physician sample has been previously described in detail.³⁴

Participating physicians were visited during 2 days in which they were providing outpatient care. Consecutive patients seen during the 2 days of observation of each physician were invited to participate in the study. Patients agreeing by verbal informed consent were included. Informed consent was obtained from the parent when the adolescent was younger than 18 years and from patients themselves when they were 18 years or older. Trained research nurses accompanied physicians in the examination room and directly observed consecutive patient encounters. Physicians and patients were blinded to specific study hypotheses. Based on the criteria used by GAPS, patients between the ages of 11 and 21 years were identified as adolescents and included as subjects for this study.

Multiple methods of data collection were used. Trained research nurses directly observed the patient-physician encounter. They measured the content of the visit using a postobservation checklist that documented services and examinations performed during the visit and a number of other visit characteristics. A patient exit questionnaire was completed after each visit by the adolescents with help from their parent or guardian, if needed. The patient questionnaire measured demographic characteristics (including patient age, sex, and insurance information), health status, patient satisfaction, and patient report of services that were delivered during the visit. Subsequent to the days of observation, nurses conducted a medical record review of the patients who were directly observed. Physicians completed a questionnaire following their participation in the study. The physician questionnaire measured a number of factors, including demographic characteristics of the physician (age, sex, marital status, years in practice, and residency training), knowledge of the US Preventive Services Task Force guidelines, satisfaction, health status, and personal health habits. Research nurses also completed a practice environment checklist measuring practice characteristics following their days of observation. Data specifically concerning preventive health–habit counseling during each visit were collected using direct observation of the patient visit by a research nurse and medical record review of services delivered during the observed visit and all visits within the previous year.

The preventive services of interest for this study were the health-habit counseling services recommended annually by GAPS for all adolescents, regardless of risk. They included counseling on diet; fitness; lifestyle (including sexual behavior and tobacco, alcohol, and drug use) and injury prevention; and screening for eating disorders and use of tobacco, alcohol, or drugs.¹ The delivery of specific preventive services during the observed visit was ascertained from the direct observation checklist of performed services.

The main outcome of interest for this study was whether or not any health-habit counseling service was delivered to the adolescent during the observed visit. In addition, 2 secondary outcomes were used to describe which services were being delivered. First, the rate of delivery for each GAPS-recommended counseling service was calculated as the number of adolescents who received the service according to direct observation, divided by the total number of adolescents. Second, the proportion of adolescents up to date on each preventive counseling service was determined. Adolescents were considered up to date for a particular service if they received the service at the observed visit or within 1 year according to the medical record.

Data on patient demographics, including age, sex, whether they were a new patient, and the number of visits within the past year, were obtained from the medical record. The patient's type of insurance was ascertained from billing data and verified by the patient's exit questionnaire. Race, classified as white or nonwhite, was assessed by direct observation by the research nurse. Health status was measured on the patient's exit questionnaire using a modified version³¹ of the Medical Outcomes Study 6-item health survey.³⁵ It was measured using a 5-point Likert-type scale, with 1 indicating poor health and 5 indicating excellent health. Visit characteristics, including the type of visit, length of visit, presence of another family member, and discussion of other family member's problem during the visit, were measured by direct observation. Whether or not any prevention flow sheet was used by the physician was abstracted from the medical record. Physician characteristics, including age and sex of the physician and familiarity with the US Preventive Services Task Force guidelines, which are similar to GAPS recommendations, were obtained from the physician questionnaire. The type of the practice (solo vs group) and the location (urban vs rural) were obtained from the practice environment checklist.

Descriptive statistics on characteristics of the patients, visits, physicians, and practices were calculated. The major focus of the analysis was to explore which patient, visit, and physician characteristics were associated with the delivery of any counseling service during the visit. Univariate analyses of dichotomous or categorical predictors used χ² tests, and continuous data used Student t tests.

Multivariable logistic regression was used to model the relationship between the patient, visit, and physician predictors and the delivery of preventive services. The predictors that were included were those that were univariately associated with preventive counseling service delivery (P<.05). Patient age and sex were also included in these models because they were hypothesized as being contextually important a priori.

Data were collected on 4454 visits by patients of all ages. Of the 4454 total visits observed, 10% were made by adolescents. These 445 adolescent visits represent the sample for this study. Of the 138 physicians participating in the larger study, 119 (86.2%) saw adolescent patients during the 2 different days of observation, ranging from 1 to 13 adolescent patients per physician. Participating physicians were demographically similar to active practicing members of the American Academy of Family Physicians³⁶ in mean age (42.8 vs 45 years) and self-reported number of patients seen per week (108.3 vs 106). They differed only by having a higher percentage of female physicians (31.2% vs 18%) and residency-trained physicians (89% vs 68%), which reflects recent trends in the characteristics of family physicians.

The demographic characteristics of the sample are shown in Table 1. Of the 445 subjects, 59.1% were female. The mean health status score was 4.0 on a 5-point scale, indicating that, on average, the adolescent subjects reported being in good health. Most of the sample (77.1%) had visited their physician 1 or more times within the year before the observed visit. The mean was 3.5 previous visits (range, 1-9). Most of the observed visits (73%) were for the treatment or follow-up of an acute illness. The mean visit length was 8.6 minutes, with the longest visit lasting nearly 50 minutes. Adolescents were accompanied by another family member for at least a portion of the patient-physician encounter during 61% of visits.

The rate of delivery for each specific counseling service during the observed visit is shown in the first 2 columns of Table 2, stratified by the reason for visit. The rate of delivery for most counseling services was higher during well-care visits than during illness visits. The rate of health-habit counseling during both types of visits, however, was low. Advice on exercise was the single most frequently delivered counseling service during both well-care and illness visits (20% and 11.4%, respectively). Counseling on at least 1 of any of the diet issues occurred in 20% of well-care visits, counseling on at least 1 accident prevention item occurred during 21.7% of well-care visits, and counseling on any contraception issue occurred during 25% of well-care visits.

In addition to the rate of delivery of preventive services at the current visit, the rate at which adolescents were up to date (having received the service within the past year) is shown in the third column of Table 2. Even for the most commonly delivered health-habit counseling—exercise—only about one fifth of the adolescents in the sample were up to date, as measured by direct observation of the index visit and a review of medical records for the previous year.

For analyses of factors associated with the delivery of health-habit counseling, the outcome measure was whether or not any GAPS-recommended counseling service was delivered during the visit. Of all observed visits, 37.5% included the delivery of at least 1 counseling service. Even among designated well-care visits, 40% included no health-habit counseling. Univariate analyses of the outcome with patient, visit, physician, and practice characteristics are shown in Table 3. Adolescents receiving 1 or more preventive services were on average older and were more likely to be a new patient (P<.001 for both) than those not receiving preventive services. No difference was found in sex, self-reported health status, type of insurance, or number of previous visits within the past year.

Numerous visit characteristics were associated with preventive counseling service delivery. Specifically, visits including preventive services were on the average 2.5 minutes longer, and they were more likely to be well-care visits (P<.001). The presence of another family member during the visit was associated with a decreased delivery of health-habit counseling services (P<.001).

Multivariable logistic regression was used to model the relationship between the delivery of preventive counseling and patient, visit, physician, and practice characteristics that were significant in univariate analyses. The variable concerning the discussion of another family member's problem during the visit was not included in the multivariable analyses because it was highly correlated with the presence of another family member. The remaining univariately associated factors, plus patient sex, was entered into the multivariable model. The coefficients, odds ratios, and 95% confidence intervals associated with each of the factors included in the final model are shown in Table 4.

Patient age was the factor most strongly associated with preventive counseling service delivery at the multivariable level (χ²₁=15.7; P<.001). The odds ratio associated with patient age was 1.22, indicating a 22% increase in the odds of receiving preventive services with each additional year of age. The other variables that remained significantly associated with preventive counseling service delivery at the multivariable level were length and type of visit and whether the patient was a new or established patient. A χ² goodness-of-fit test indicated that the model significantly aided in predicting preventive counseling service delivery (χ²₁₀=82.6; P<.001), correctly predicting preventive service delivery in 69% of the visits.

As the rate of risky health behaviors during adolescence continues to rise,² the provision of preventive services becomes increasingly important.³⁷ Understanding the patterns of prevention and having accurate estimates of preventive service delivery rates to adolescents in primary care are critical in designing interventions for clinicians to improve prevention rates and ensure that these young people are receiving the services they need. Furthermore, knowing which services are and are not routinely performed may be important for health policy decisions and educational curricula concerning this age group.

In this study, the overall rate of delivery of health-habit counseling preventive services during adolescent visits was low. Only 37.5% of all visits included a GAPS-recommended counseling service. Even when taking into account multiple visits during the year using the up-to-date measure, the rate at which these persons receive counseling services was far below the recommendations. Due to the use of the medical record in measuring the "up to date" outcome, however, and the known underreporting of health-habit counseling,³¹ the reported percentage of adolescents who were up to date on each preventive counseling service represents a conservative estimate, and actual rates may be higher.

One disturbing finding of the study is that 40% of designated well-care visits included no GAPS-recommended health-habit counseling service. This clearly represents a missed opportunity for preventive counseling. The implication is that, without a change in current practice, increasing prevention simply through scheduling more well-care visits may not be adequate. Individual compliance with annual well-care visits is likely to be low because well-care visits are not always reimbursable by insurance policies and the salience of preventive care for adolescents is low based on clinical experience and the fact that most adolescent visits are for the treatment of an illness, as found in this study. In addition, illness visits may represent the "teachable moments" for certain health-habit counseling.³⁸ Therefore, effective strategies should be developed to promote the delivery of preventive services opportunistically during visits for acute or chronic illnesses, in addition to efforts to foster well-care visits for teenagers and improve delivery within these visits.

Despite the low overall rates of preventive counseling service delivery, clearly some counseling services are delivered more often than others. The discrepancies in service-specific delivery rates indicate that there is not 1 universal barrier to delivery that prohibits all preventive services from being delivered.

Physicians may be less likely to provide those preventive services that encompass personal or sensitive issues.^15,22 Steiner and Gest³⁹ recently found that although, during any given visit, many young people may not want to discuss their health-risk behaviors with their physician, receptivity to such a discussion increased when the physician brought up the issues at several visits. Furthermore, most young people think it is the physician's responsibility to discuss health-risk behaviors with them.^39- 40 Despite the report that adolescents may be willing to have discussions with their physicians about sexual activity and substance use, these activities are rarely discussed, as evidenced by the data from this study. Efforts are needed to decrease the social stigma of these discussions.

The study findings, like the findings of others,²⁴ indicate that older patient age is strongly associated with increased delivery of counseling preventive services. However, GAPS recommends that each of these counseling preventive services be delivered annually to all persons aged 11 to 21 years. This recommendation is substantiated by recent results from the Youth Risk Behavior Survey, which finds that adolescents are engaging in health-risk behaviors at increasingly younger ages.²

Another important finding from this study is the strong association between the length of visit and the delivery of counseling preventive services. The visits in which counseling preventive services were delivered were on average 2.5 minutes longer than those visits in which no counseling was done. This represents an important finding because a lack of time is commonly cited^38,41- 44 by primary care physicians as a substantial barrier to the provision of preventive services. Time constraints may be seen within the context of the competing demands of the primary care visit.⁴⁴ Primary care physicians are responsible for addressing illness, psychosocial and administrative issues, and prevention. Competing priorities represent a substantial barrier for young people and their families in seeking preventive care.⁴⁵

The finding that new patients were more likely to receive counseling preventive services than established patients is contrary to the notion that continuity of care increases the delivery of preventive counseling service.⁴⁶ A possible explanation for this finding is that physicians routinely incorporate certain preventive services into their first visit with new patients, such as the ascertainment of tobacco or alcohol use. Asking about risky health behavior may be seen as routine during new patient visits, but such questioning is more stigmatized if the physician and adolescent patient already know each other.

The strong univariate association between the presence of another family member during the visit and a decrease in the delivery of preventive counseling service has implications on clinical practice because it underscores the need for confidential health care for adolescents.^37,47 Confidentiality is extremely important in adolescent health. Previous studies^22- 23,40 have found that a perceived lack of confidentiality is a barrier for young people in seeking health care. A recent study⁴⁸ found that assurances of confidentiality by the physician were associated with an increased willingness by young people to disclose sensitive information. Therefore, asking parents to leave for a portion of the visit is a vital strategy to increase the physician's ability to openly discuss risky health behaviors with this age group.⁴⁷

An additional barrier to health-habit counseling delivery to young people may be a lack of acceptance of certain GAPS recommendations on the part of family practitioners. Indeed, a recent survey^49- 50 of physician agreement with preventive service guidelines found that family physicians disagreed with recommendations for health-habit counseling around sensitive behavior issues such as sexual behavior and substance abuse. Although some⁵¹ have criticized the GAPS recommendations for being based less on evidence than on strongly held opinions by members of the panel responsible for GAPS, the GAPS recommendations are similar to those of the US Preventive Services Task Force¹⁶ and the American Academy of Family Physicians,¹⁷ and these findings are likely to apply to family physicians' acceptance of these recommendations as well.

The rate of delivery of health-habit counseling to adolescents seen in community family practice is far below current recommendations. This raises concern about the feasibility of existing preventive service recommendations for this age group, given the competing demands of community primary care practice. Yet, the GAPS recommendation for annual well-care visits is an ideal worth striving toward. Current limitations in access to primary care, the limited salience of prevention to many young people, and the emphasis in primary practice on caring for illness all work against achieving the ideal in the current health care climate. Strides toward this goal may need to be made incrementally.

Efforts to increase the rate of the delivery of counseling preventive services to young people should focus on scheduling dedicated visits for well care and focusing these visits to a greater degree on health-habit counseling, using illness visits as opportunities for preventive service delivery, and asking accompanying family members to leave for at least a portion of the visit to offer the patient confidentiality to discuss sensitive issues.

Accepted for publication September 1, 1998.

This research was supported by grant 1RO1 CA 60862 from the National Cancer Institute, Bethesda, Md, and by a Generalist Physician Faculty Scholars Award (Dr Stange) from Robert Wood Johnson Foundation, Princeton, NJ.

We thank the physician members of the Research Association of Practicing Physicians and the office staffs and patients, whose participation in this study made it possible.

Reprints: Meredith A. Goodwin, MS, Department of Family Medicine, Case Western Reserve University, 10900 Euclid Ave, Cleveland, OH 44106-7136 (e-mail: mag@po.cwru.edu).

Editor's Note: Direct observation showed that the family practitioners in this study didn't do so well. Any bets on how pediatricians would fare? So who will do the study?—Catherine D. DeAngelis, MD