Improved Preventive Care for Asthma: Title and subTitle BreakA Randomized Trial of Clinician Prompting in Pediatric Offices FREE

Asthma is the most common chronic illness of childhood,¹ and morbidity has increased despite improvements in therapy.² National guidelines for the management of asthma are available,^{3 - 5} and use of these guidelines can reduce morbidity and mortality due to asthma.^{5 - 10} However, several studies have documented poor physician adherence to the guidelines,^{11 - 15} likely contributing to otherwise preventable illness. Children from poor or minority backgrounds suffer the greatest morbidity from asthma^{16 - 18} and are the least likely to receive appropriate preventive treatment.^{19 - 21}

An office visit provides an opportunity for the clinician to apply asthma guidelines to clinical care.⁴ Unfortunately, during many patient encounters, opportunities to provide guideline-based preventive care are missed. We know, for example, that many providers fail to document symptom severity in the medical record²² and underestimate the severity of their patient's symptoms.^{13 ,23} Similarly, written action plans^{22 ,24} and asthma education are not provided consistently.²⁴ Several barriers associated with poor implementation of the guidelines have been documented.^{11 ,25}

Provider prompts can be effective in stimulating action regarding clinical issues and improving preventive care.^{26 - 28} The objective of this study was to determine whether clinician prompting regarding asthma symptom severity and guideline recommendations at the time of an office visit improves preventive care delivery for inner-city children with persistent asthma.

We conducted this study in 2 inner-city pediatric practices in Rochester, NY, from November 20, 2003, through September 14, 2005. The practices are training sites for the residency program at the University of Rochester, are staffed by attending physicians and nurse practitioners, and provide care primarily for low-income, minority patients. Children aged 2 to 12 years with at least 2 asthma exacerbations in the preceding 2 years were identified as potential study subjects. We reviewed medical charts daily for children with scheduled visits for well-child care, asthma care, or non–asthma-related illness care. A study team member approached parents of identified children in the waiting room before their visit with the clinician and completed a detailed survey to assess eligibility. The institutional review board of the University of Rochester approved the study protocol.

Eligibility required children to have symptoms consistent with mild persistent or more severe asthma⁵ and thus to need enhanced preventive care. Parents were asked the following specific questions about their child's current asthma symptoms: (1) During the past 2 weeks, how often has your child had coughing, wheezing, or shortness of breath during the day (less than twice a week, twice a week, more than twice a week but not every day, everyday but not all the time, or everyday/all the time)? and (2) During the past 2 weeks, how often has your child had coughing, wheezing, or shortness of breath during the night (never, once every 2 weeks or less, once a week, more than 1 night a week, or frequently/every night)? Children of parents who indicated daytime symptoms of at least more than twice a week but not everyday or nighttime symptoms at least once a week were considered to have persistent asthma and were eligible to participate. Because of a lack of available resources, Spanish-speaking families were enrolled into the study only on days when a Spanish-speaking interviewer was available. Eligible children had to be accompanied by a parent or legal guardian. We excluded children with other medical illnesses that would complicate the assessment of asthma symptoms (eg, cystic fibrosis or heart disease).

After written informed consent was obtained, a baseline survey was conducted with the parent to obtain information regarding household demographic characteristics, asthma symptom severity, medication use, and environmental tobacco smoke (ETS) exposure.

After completion of the baseline assessment, each child was assigned randomly to the clinician-prompting group or the standard-care group. Randomization was stratified by office site and preventive asthma medication use at baseline and was blocked in groups of 6 to ensure an equal number of subjects in each group for each seasonal time period. Only 1 child was randomly selected for enrollment from eligible sibling pairs who arrived together for an office visit. Randomization cards were made from a table of random numbers and were kept in sealed, opaque, sequentially numbered envelopes until after the baseline assessment.

The study team member enrolling the child in the waiting room prepared an asthma report and action recommendation form to serve as the prompt for each child in the clinician-prompting group (Figure 1). This prompt included information regarding the child's current asthma symptoms, as obtained from the baseline survey conducted in the waiting room, the classification of the child's asthma severity, and recommendations for treatment based on the national guidelines.⁵ The prompt also included information regarding the child's exposure to ETS, with recommendations for discussion of ETS reduction.

The prompt was given to the parent in the waiting room, and the parent was instructed to hand it to the clinician when the child was called back for the office visit. We also included a blank asthma action plan form along with the prompt for the health care provider to complete and return to the family at the end of the visit. Parents were specifically instructed to hand these forms to the clinician in the examination room, rather than to the nurse or the aide who might be initiating the visit with the child.

Parents of children in the standard-care group completed the baseline assessment in the waiting room, but no asthma prompt was created, and no information regarding the interview was shared with the provider. After the baseline assessment, the office visit proceeded according to usual care.

Before the initiation of the study, the principal investigator (J.S.H.) met with health care providers in both offices to describe the intervention and orient them to the prompt. Each practice was given a donation of $500 to thank them for their participation.

Interviewers blinded to the child's group assignment called the parent or guardian who accompanied the child to the office visit to assess the outcome measures for the study. To optimize recall, we excluded interview follow-up data for children who did not complete a follow-up within 60 days of the intervention or who had another office visit between the time of enrollment and follow-up (n = 10 in the clinician-prompting group and n = 10 in the standard-care group). The follow-up data included in the analyses were collected, on average, 11 days after the visit (range, 0-57 days). We asked parents to think about the office visit and answer specific questions about what occurred during the visit. The interview included inquiries regarding whether the clinician discussed asthma symptoms, emergency visits and absenteeism due to asthma, medication compliance, environmental triggers, and how to properly use medications. We also asked whether the clinician recommended a specific asthma follow-up visit or a visit with an asthma or allergy specialist, and whether parents were given a written asthma action plan. We inquired about whether the child received a new preventive medication different from the one he or she was already taking or if a dosage of a preventive medication he or she was already taking was changed. Finally, we asked parents whether anything else was done at the office visit to help their child's asthma.

We also performed blinded reviews of the children's medical records from that office visit to assess these same preventive measures on the basis of the health care provider's report. A study team member used a standardized chart abstraction form to collect data from the medical record, and any documentation of asthma care was recorded. The principal investigator (J.S.H.) performed a second blinded review of all medical records and completed medical chart abstraction forms to ensure accuracy of the medical-related data.

Preventive actions included provision of an asthma action plan, treatment of a comorbid condition (eg, allergies, sinusitis, or gastroesophageal reflux), recommendation of a follow-up visit for asthma, referral to a specialist, discussions related to asthma (inquiries regarding emergency visits and absenteeism due to asthma, environmental changes, and adherence to preventive medications), and changes in the preventive medication (prescription of a new preventive medication or increase of the dose of a current preventive medication). We defined smoke reduction counseling as any discussion regarding smoke avoidance or smoking cessation or education regarding the harm of ETS exposure for those households where 1 or more members smoked.

The primary outcome for this study was any preventive action occurring during the study visit (defined as any of the actions listed in the preceding paragraph). If the parent or the medical record indicated the occurrence of a preventive action, we assumed for the purposes of this study that an action had been taken. The parent interview and medical chart review data yielded the same result for this primary outcome measure in 75% of cases.

Independent variables included sex, age, race (white, African American, or other), ethnicity (Hispanic or not Hispanic), Medicaid insurance (yes or no), caretaker education (less than high school or high school or more), asthma severity (mild persistent, moderate persistent, or severe persistent), type of provider conducting the visit (attending physician, nurse practitioner, or resident), whether the visit was conducted by the child's usual health care provider (yes or no), and the number of smokers living in the home (0 or ≥1).

All randomized subjects were kept in their originally assigned treatment groups for analysis. Analysis was performed using commercially available statistical software (SPSS, version 14.0; SPSS Inc, Chicago, Ill). We used χ² analyses to compare the dependent variables between the 2 groups. A single post hoc analysis was performed with data stratified by the type of office visit (asthma visit, well-child visit, or non–asthma-related illness visit). We also performed multiple logistic regression analyses to predict the key outcome variables after controlling for the child's age, sex, race, ethnicity, and asthma severity; a smoker living in the home; and the provider's characteristics.

In planning this study, the sample size calculations were based on the ability to detect a change of 20% in the primary outcome variable (any preventive action) (improvement from a baseline of action being taken in 56% of visits²⁹ to occurrence in 76% of visits). A total sample of 200 patients was required (α = .05; β = 0.20; 2-sided test).

We reviewed the medical charts of 722 children who visited the clinics and approached 510 (70.6%) of these subjects, who were potentially eligible. Of the 510 approached, 186 were ineligible for reasons such as lack of an asthma diagnosis according to the parent, mild asthma symptoms, or lack of Spanish-speaking enrollers (Figure 2). Among the 324 eligible children, 78 were not enrolled because of parent refusal (n = 11) or lack of time in the waiting room to complete enrollment before the office visit began (n = 67). The remaining 246 children were enrolled in the program for a response rate of 75.9%. One hundred twenty-two children were randomized into the clinician-prompting group and 124 were randomized into the standard-care group. After randomization, 20 subjects were determined to be ineligible (n = 10 in the clinician-prompting group and n = 10 in the standard-care group) and were removed from further analysis (Figure 2). Seventy-nine different health care providers participated in the trial and 37 had patients in each treatment arm.

Two hundred twenty-six children had data available for analysis for the primary outcome (n = 112 in the clinician-prompting group and n = 114 in the standard-care group). We were able to include data for a subject if an eligible follow-up was completed with the parent or if a medical record abstraction was available. The overall response rate for parent follow-up interviews was 78.3% (78.5% in the clinician-prompting group and 78.1% in the standard-care group). We reviewed the medical charts for all 226 subjects.

Table 1 describes the demographic characteristics for the children and caregivers. Most of the children were male (58.0%), African American (64.2%), and insured by Medicaid (75.2%). Overall, 51.3% of the children were aged 2 to 5 years and 29.6% were Hispanic. Most of the caretakers had at least a high school education (62.8%). Among the children, 59.3% were using a preventive medication at baseline and 38.1% had severe persistent symptoms. On the day of the studied visit, 26.5% of the children were presenting to the office for asthma-related care, 33.2% for well-child care, and 40.3% for care for an illness other than asthma. Approximately half of the children (48.2%) saw their usual health care provider at the visit.

After randomization, there were fewer boys in the clinician-prompting group compared with the standard-care group (49.1% vs 66.7%). In addition, a smaller proportion of children in the clinician-prompting group lived with 1 or more smokers compared with children in the standard-care group (35.7% vs 57.0%). There were no significant differences in other baseline measures between the 2 groups.

Table 2 shows the overall results of the intervention for the 226 children included in the analysis. The children in the clinician-prompting group were more likely to have any preventive action related to asthma taken at the visit compared with children in the standard-care group (86.6% vs 69.3%; P = .002). Specifically, visits for children in the clinician-prompting group were more likely to include the delivery of an action plan (50.0% vs 23.7%; P<.001), recommendation for a specific asthma follow-up visit (53.6% vs 36.8%; P = .02), and discussions regarding asthma (75.0% vs 63.2%; P = .05) compared with visits for children in the standard-care group. Among children with 1 or more smokers in the home (n = 105), those in the clinician-prompting group received smoke reduction counseling more frequently than those in the standard-care group (57.5% vs 35.4%; P = .04). There were no statistical differences in referrals for specialist care, treatment of comorbid conditions, or changes in preventive medication between the 2 groups.

Table 3 shows the results of the intervention stratified by type of office visit. Among the 60 children presenting for an asthma-related visit, there were no differences between the clinician-prompting and the standard-care groups, and preventive actions were taken for most of the children. The greatest effect of the intervention appeared to be for the 91 children presenting to the office for non–asthma-related illness care. For these children, 81.4% of those in the clinician-prompting group received any preventive action related to asthma, compared with 54.2% in the standard-care group (P = .007). The intervention also seemed to have some effect on those children presenting for well-child care, with more than twice as many children in the clinician-prompting group receiving an asthma action plan compared with children in the standard-care group (57.9% vs 27%; P<.01).

Almost all of the significant findings from the overall analysis remained statistically significant in multivariate regression analyses after we controlled for baseline asthma severity, provider characteristics, child demographic factors (age, sex, race, and ethnicity), and exposure to ETS (Table 4). In these analyses, children in the clinician-prompting group had 3-fold greater odds of having any preventive action taken at the visit compared with children in the standard-care group (odds ratio, 3.1; 95% confidence interval, 1.4-6.6).

This study demonstrates that prompting clinicians about asthma severity and care guidelines at the time of an office visit results in improved preventive care delivery for inner-city children. Visits for the children in the clinician-prompting group were more likely to include discussions of asthma prevention, recommendations for specific asthma follow-up, and the delivery of an action plan at the time of the visit compared with visits for children receiving standard care. Prompting also improved smoke reduction counseling for patients whose households included a smoker. The benefits of the intervention were seen primarily among children presenting for non–asthma-related illness care or for well-child care because most children presenting for asthma care received preventive actions regardless of their group allocation.

Although the intervention improved the delivery of preventive care, one could argue that a large percentage of the children in the clinician-prompting group still had no follow-up visit recommended (46.4%), received no action plan (50.0%), and had no discussions related to asthma (25.0%). In addition, we found no difference between groups with regard to a change in the prescription of preventive medications. In part, this may be because our symptom inquiry was restricted to the 2 weeks preceding the visit; thus, some providers may have interpreted the child's symptoms as a manifestation of viral respiratory infection rather than of persistent asthma. This would not necessarily warrant guideline-based preventive care.³⁰ The prompt was designed to serve as a screen for current persistent symptoms, not as a definitive diagnostic measure. In addition, because we considered actions taken during only 1 visit, it is possible that patients received a preventive intervention at a different visit.

Several studies^{31 - 34} have evaluated other methods to improve preventive care delivery to inner-city children with asthma in the office setting. A few studies^{23 ,32 ,35} have suggested that physician education regarding asthma management and quality improvement can improve adherence to care guidelines. However, Homer et al³⁶ recently evaluated a quality improvement educational intervention for asthma in primary care offices and found no significant effect on care for children with asthma. Lozano et al³⁴ evaluated a multilevel intervention involving clinician education regarding care guidelines combined with nurse-led case management. That combined intervention was effective in improving asthma care for subjects in the intervention group, including greater adherence to controller medications. However the clinician-education component of the intervention was not sufficient to yield significant benefit for the intervention group; the additional resources provided by a trained nurse clinician for individual case management were required. Novel methods to improve asthma care within the time demands and financial constraints of a primary care office are needed.

The concept of missed opportunities at an office visit implies that individuals eligible for a particular health-related intervention are seen for care but do not receive that particular intervention. Studies have focused on identifying and reducing missed opportunities during office visits for routine immunizations,^{37 - 38} influenza immunization,^{39 - 41} and screening and counseling for health risks.^{42 - 44} Missed opportunities to improve asthma care are common^{22 ,24} and, if exploited properly, can enhance preventive care within the context of the child's usual place for medical care.

The institution of symptom screening and prompting used in the present study is relatively simple and inexpensive and yielded a significant increase in the delivery of preventive care for children with persistent asthma. This type of intervention could be implemented into routine practice with the use of electronic medical records, which are increasingly available in many pediatric offices. On their arrival for an office visit, patients with asthma could be identified by codes from the International Classification of Diseases, Ninth Revision, and parents could enter the current symptom information into computerized templates. This information could be translated electronically into a prompt for the clinician as a computerized printout or a prompt on the screen of an examination room computer. Alternatively, in the absence of electronic medical records, one could design a manual system by using written forms and medical charts to establish symptoms, severity classification, and guideline recommendations. Forms could be given to the parent at the time he or she checks in for an appointment, completed in the waiting room by the parent, and delivered to the health care provider at the time of the visit. The translation of symptom information into the guideline recommendations would be facilitated by appropriately designed screening forms with simple algorithms for the provider. Such changes to the system of care would allow the clinician to avoid missed opportunities and to provide guideline-based asthma care within the context of a busy practice.

There are some potential limitations to this study. Blinding of parents and clinicians to the allocated treatment groups was not possible, and self-reported responses could have been biased. Furthermore, although the study team specifically asked the parents in the clinician-prompting group to give the prompt to their health care provider during the visit, this did not occur in all cases and thus may have led to an underestimation of the intervention effect.

Although physician education was not a key feature of the intervention, clinicians were oriented to the prompt and reminded of care guidelines before the start of the study. Because randomization was on the patient level rather than the clinician level, clinicians of patients in both treatment arms received this orientation. This may have served as a reminder to the clinicians to use guideline-based care and thus may have created a conservative bias, making differences between the 2 treatment groups more difficult to detect.

Medical chart review data often lack details of the clinical encounter^{22 ,45} ; thus, we may have missed some preventive care actions that occurred. However, we also used parent-reported data, which is more accurate in determining what happens at a visit.⁴⁵ The clinicians for children in the intervention group received a prompt during only 1 visit, and the prompt included only general recommendations based on the guidelines. Prompting at multiple visits with detailed recommendations tailored to individual patients (eg, suggesting specialist referrals for children with severe persistent asthma) may have yielded even more significant findings with improvements across additional areas of preventive care. Finally, only 2 urban clinics in Rochester were included in the study; thus, the findings can be generalized only to similar practice sites.

This study demonstrates improved delivery of preventive care for urban children with asthma with the use of symptom-based clinician prompting in primary care offices. Because guideline-based care can improve asthma outcomes,^{5 - 10} the potential implications of these findings are great. In addition, because urban children are at highest risk for receiving suboptimal care, incorporating such a method of systematic screening and prompting into routine visits for these children may substantially reduce asthma morbidity.

Correspondence: Jill S. Halterman, MD, MPH, Department of Pediatrics, University of Rochester School of Medicine and Dentistry, PO Box 777, 601 Elmwood Ave, Rochester, NY 14642 (jill_halterman@urmc.rochester.edu).

Accepted for Publication: May 27, 2006.

Author Contributions: Dr Halterman had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Halterman, Fisher, and Szilagyi. Acquisition of data: Halterman, Conn, Fagnano, Lynch, and Marky. Analysis and interpretation of data: Halterman, Fisher, Conn, and Szilagyi. Drafting of the manuscript: Halterman, Fagnano, and Lynch. Critical revision of the manuscript for important intellectual content: Halterman, Fisher, Conn, Marky, and Szilagyi. Statistical analysis: Fisher and Conn. Obtained funding: Halterman. Administrative, technical, and material support: Conn, Fagnano, and Lynch. Study supervision: Szilagyi.

Funding/Support: The study was supported by grants from the Halcyon Hill Foundation and the Robert Wood Johnson Foundation's Generalist Physician Faculty Scholars Program.

Acknowledgment: We thank George B. Segel, MD, for his insightful review of the manuscript.