Objectives. Follow-up with a primary care provider (PCP) is recommended after an emergency department (ED) visit for asthma to assess clinical status and develop a management plan to improve future care. However, previous ED-based studies of urban children with asthma have reported low follow-up rates. The objective of this study was to determine whether scheduling an appointment at the time of an ED visit improves PCP follow-up for urban children. A secondary goal was to assess the effect of this intervention on short-term health outcomes and the use of recommended preventive controller medications.
Methods. This randomized trial enrolled a convenience sample of children who were 2 to 18 years old and discharged after treatment for acute asthma in an urban children’s hospital ED. Both intervention and control subjects were instructed to follow up with their PCP within 3 to 5 days. Study staff assisted intervention subjects to call their PCP from the ED and schedule an appointment. When follow-up could not be scheduled, assistance continued after ED discharge by telephone until an appointment date was confirmed. Study outcomes included PCP visits, asthma-related morbidity, and daily use of preventive medication 4 weeks after the ED visit. Outcomes were assessed by telephone interview and confirmed by PCP record review.
Results. A total of 278 eligible subjects were enrolled over 8 months; intervention and control groups were similar by demographic variables and PCP type as well as by asthma history, symptoms, and previous medication use. Only 38% of subjects reported using a daily controller medication, although 70% described persistent asthma symptoms for which these are recommended. For the intervention group, follow-up appointments were successfully obtained during the ED visit for 24% of subjects; when unsuccessful, a median of 3 telephone calls (range: 1–14) were needed to confirm that an appointment had been scheduled. During the 4 weeks after the ED visit, intervention subjects were more likely than controls to follow up with their PCP (64% vs 46%; relative probability for follow-up: 1.4; 95% confidence interval: 1.1–1.7). Study groups did not differ in return ED visits, missed school or work, or the percentage reporting daily use of a controller medication (58% vs 54%) 4 weeks after the ED visit. The median time to the next PCP visit was shorter among intervention subjects (13 vs 54 days).
Conclusions. Scheduling an appointment after an ED visit increased the likelihood that urban children with asthma would follow up with a PCP. An appointment could not be obtained during the ED visit for most children. Other interventions are needed to improve linkage between ED and primary care for asthma and to improve the use of controller medications.
Efforts to improve care for childhood asthma have received increasing attention in recent decades. A marked rise in asthma prevalence since 19801 as well as increases in asthma-related mortality,2 hospitalizations,1,2 disability,3 and cost4 have led to the development of practice guidelines such as the National Asthma Education and Prevention Program (NAEPP): Guidelines for the Diagnosis and Management of Asthma, published by the National Heart, Lung and Blood Institute.5 Central to these guidelines is the importance of ongoing preventive care and follow-up with a primary care provider (PCP). The efficacy of preventive care in improving outcomes such as emergency department (ED) visits and hospitalizations has been supported by evidence from multiple studies.6–9
The ED is an important source of acute care for children with asthma, accounting for more than half a million visits in 1995.1 Because preventive care has been shown to reduce future ED asthma visits, follow-up with a PCP is particularly important after ED discharge to reassess the management plan; the NAEPP Guidelines recommend follow-up within 3 to 5 days. Although the recommendations are clear, research has found poor adherence to these guidelines, particularly in urban populations at high risk for asthma morbidity.10–12 Researchers from the National Cooperative Inner City Asthma Study interviewed 334 inner-city families and estimated that only 52% saw a PCP within 1 month of an ED asthma visit.13 Other results from this study suggest that difficulties in getting appointments and gaining access to an existing PCP were barriers to follow-up frequently cited by these inner-city families.14
This study was designed to assess the efficacy and feasibility of providing a PCP follow-up appointment after an ED visit for asthma. Using a randomized, controlled trial design, we used patient interviews and PCP medical record review to assess whether this intervention would 1) increase the rate of PCP follow-up after an ED visit and 2) improve short-term asthma-related outcomes and the use of recommended preventive medications.
Before the initiation of enrollment, the institutional review board approved the design of the trial. Study staff were present in the ED between the hours of 8 am and midnight, 7 days per week. Families were approached for enrollment on the basis of the following inclusion criteria: 1) age 2 to 18 years, 2) residing within Philadelphia city limits, 3) history of asthma defined as 2 or more previous medical visits at which bronchodilators were prescribed, 4) acute respiratory symptoms requiring treatment with bronchodilators in the ED, and 5) a decision to discharge the patient had been made by the treating physician. Children with a history of cystic fibrosis or congenital cardiac disease were excluded. Because of the need for ED and telephone interviews and consent for PCP medical record review, patients were excluded when no English-speaking guardian was present or when the family had no way to be reached for telephone follow-up. A patient log was kept during the study period to document all patients screened for the study and the reasons for exclusion. Patients were not enrolled in the study more than once.
During the ED visit, study staff surveyed subjects about demographic characteristics and asthma history and treatment. Using categories defined by the NAEPP Guidelines, subjects were categorized as having persistent asthma when they reported baseline symptoms (cough, wheeze, shortness of breath, trouble breathing, or chest tightness) more than twice weekly for daytime symptoms or twice monthly for nighttime symptoms. Daily preventive controller medication use was defined as the daily use of inhaled corticosteroids, cromolyn, nedocromil, or an antileukotriene agent. After obtaining written informed consent, study staff assigned patients to intervention or control groups using a series of numbered, opaque envelopes that had been prerandomized in blocks of 8. Control group patients received standard discharge instructions to follow-up with their PCP within 3 to 5 days. For intervention group subjects, study staff brought the guardian to a telephone and together attempted to contact the PCP and schedule a follow-up appointment. A second attempt was made when the office could not be reached initially. When an appointment could not be scheduled after 2 attempts, the guardian was informed that study staff would be calling them to assist with obtaining an appointment. For these patients, staff called every 1 to 2 days until an appointment had been scheduled or 2 weeks had elapsed since the ED visit. When the family could be reached during daytime hours, the PCP office was dialed in using a conference call function; otherwise, study staff offered to make an appointment or call to check that the family had been successful. Copies of ED medical records for both control and intervention groups were faxed to PCPs, as per standard procedure in this ED separate from the study protocol. Initiating controller medications at ED discharge was not part of routine practice during the study.
Four weeks after the ED visit, all study subjects were contacted by telephone; telephone interviewers were unaware of group assignment. A standardized telephone survey inquired about the patient’s symptoms since the visit, date of follow-up, interventions that had taken place, and reasons for lack of follow-up if no visit had occurred. At the conclusion of the study, PCP offices were contacted to verify the dates of follow-up visits. A copy of the medical record for each visit was requested for review. To determine the effect of the intervention on clinical outcomes, a subgroup of patients who were followed by primary care sites within this health system were followed for 1 year. Subsequent PCP visits, ED visits, and hospitalizations within the health system were quantified using computerized information systems.
Sample size was estimated on the basis of a predicted follow-up rate among controls of 50%. Assuming 10% loss to follow-up and setting α2 = 0.05, 280 subjects were needed to provide 90% power to detect an increase of 20% in the follow-up rate. Statistical analysis was conducted using SPSS Version 10.1 (Chicago, IL) and STATA Version 7.0 (College Station, TX). Group differences were determined using χ2 for categorical variables, independent sample t test for comparison of means of continuous variables, and Mann-Whitney U test for comparison of medians. Mantel-Haenszel risk stratification was used to calculate stratum-specific relative risk for follow-up within subject subgroups. Multiple logistic regression was used to adjust for differences in all baseline variables between control and intervention groups.
Between September 1999 and April 2000, a total of 1305 children who had asthma and were between the ages of 2 and 18 were screened by study staff. Of these, exclusions were made for the following reasons: clinical decision to admit to the hospital (506) or discharge when study staff were not present (67), nonurban residence (48), no guardian present to give consent (41), not English-speaking (7), and previous study enrollment (44). Of the 592 remaining eligible subjects, 186 were not approached because multiple patients were present in the ED simultaneously; missed patients were similar to enrolled subjects by age, race, gender, and triage acuity. Of those approached, 286 (70%) subjects gave consent for enrollment (Fig 1). Randomization assigned 142 subjects to the control group and 144 to the intervention group. Six subjects were later found to be ineligible and excluded because of previous enrollment in the study (3), nonurban residence (2), and diagnosis changed from asthma during the ED visit (1). The guardians of 2 subjects (1 intervention and 1 control) initially gave consent but then asked to be withdrawn from the study during the ED visit. These subjects were excluded in accordance with the guardian’s request, and outcome data were not collected. For another 4 subjects, the treating physician changed the initial decision to discharge home after study enrollment and randomization; these subjects were included in accordance with an intention-to-treat analysis.
Baseline characteristics for intervention and control subjects are presented in Table 1. A majority of the study subjects were male (62%), black (94%), and enrolled in Medical Assistance (62%). All but 1 of the subjects reported a site for regular medical care, predominantly in primary care clinics staffed by the residents or faculty of 1 of the Philadelphia teaching hospitals (55%). When asked to recall hospital visits for asthma during the past year, a substantial proportion reported 2 or more ED visits for asthma (55%) or a hospitalization (38%). Most (70%) described symptoms at baseline consistent with the NAEPP Guidelines definition of persistent asthma, although only 38% reported using a daily controller medication that would be recommended for these symptoms. Patients with persistent symptoms reported higher rates of controller medication use than those with intermittent symptoms, although this did not reach statistical significance (41% vs 32%; P = .2). Intervention and control groups did not differ significantly in any of these baseline variables.
For the 139 intervention group subjects, study staff assisted 97 (67%) in calling their PCP from the ED to schedule an appointment. For another 34 subjects, the call was deferred at the patient’s request (eg, could not wait or knew the office was unable to schedule at this time). The remaining 8 children already had a follow-up appointment scheduled. Of the 97 calls made from the ED, 33 (34%) resulted in scheduling an appointment. For 10 of the calls, the office was unable to be contacted after 2 attempts, and for the remaining 54, the office staff instructed the patient to call back during business hours. The success of ED calls was related to the time of day, with only 2 appointments scheduled when the time of ED discharge was after 6:00 pm. For the 98 intervention subjects who were discharged from the ED without a follow-up appointment, study staff made a median number of 3 (range: 1–14) telephone calls to ascertain that an appointment had been scheduled. Appointments were scheduled for 81 of these children. An additional 4 families planned to walk in unscheduled for follow-up; 2 said that they had spoken with their PCP personally on the telephone and been told that a follow-up visit was not needed. For 11 (8%) of the intervention subjects, no plan for follow-up was able to be verified during the 2 weeks after the ED visit. The median time to the scheduled follow-up appointment was 5 days from the ED visit (range: 1–88 days).
Telephone interviews 4 to 6 weeks after the ED visit were completed for 260 (94%) subjects; completion rates were similar for control and intervention groups (Table 2). When asked about the 4 weeks after the ED visit, interviewees reported a return visit to the ED for 11% of subjects and a median of 2 days of school missed for the patient (range: 0–21) and 1 day of work for the guardian (range: 0–21). These outcomes did not differ significantly between intervention and control groups. The proportion of subjects who reported a return to baseline health was greater in the intervention group compared with controls (82% vs 71%; P = .03). A greater proportion of intervention subjects reported seeing a PCP compared with controls (77% vs 51%; P < .001). When asked for the reason for lack of follow-up with a PCP, the most common reasons given were that an appointment was not necessary (9% of all intervention subjects vs 25% of controls) or that there was difficulty in obtaining an appointment (5% vs 14%). Overall, 56% of subjects reported using a daily controller medication, and this rate was similar among intervention and control subjects (58% vs 54%; P = .5). When this analysis was limited to subjects who had described persistent symptoms at the time of the ED visit, a trend toward increased use of controller medications was observed among intervention subjects (64% vs 53%; P = .11). Overall, subjects who had persistent asthma symptoms at the initial ED visit and saw a PCP within the 4 weeks of the ED visit were more likely to report using a controller medication 4 weeks after the ED visit than those who did not make a follow-up appointment (65% vs 46%; P = .01).
Verification With Primary Care Site
Primary care sites were contacted at least 6 months after the ED visit for 250 (90%) subjects. For the remaining subjects, the provider either did not respond (9) or could not locate records for the subject (19). Overall, 23% of subjects saw a PCP for follow-up within the 5 days recommended by the NAEPP Guidelines; this rate was significantly greater among intervention subjects than controls (29% vs 17%; P = .03). The rate of PCP follow-up within 4 weeks for subjects with verification was significantly greater among intervention subjects compared with controls (63% vs 44%; P = .002). Kaplan-Meier survival analysis was used to compare time to PCP follow-up in intervention and control groups (Fig 2). A total of 29 subjects had not yet made a subsequent visit at the time of contact with the PCP, and these were treated as censored data for the purpose of this analysis. The median time to the next primary care visit was significantly less for intervention group subjects compared with controls (13 vs 54 days; P = .004), and the 2 survival curves were significantly different by log-rank analysis (P = .01).
For testing the validity of outcomes reported by telephone interview, agreement was assessed with outcomes obtained from the PCP. For the 233 subjects who had both telephone interviews and verification, agreement regarding follow-up was observed in 185 (79%; κ = 0.6). For 94 subjects for whom the medical record of the visit was available, agreement about the prescription of preventive medications was observed in 81 (86%; κ = 0.7).
By combining the telephone interview and primary care verification, an outcome was determined for all but 1 of the study subjects. For cases in which there was disagreement, the verified result was used as the final study outcome. Overall, the follow-up rate among intervention subjects was significantly greater than among controls (64% vs 46%; P = .003). The relative probability for primary care follow-up for intervention subjects compared with controls was 1.4 with a 95% confidence interval (CI) extending from 1.1 to 1.7.
For assessing the efficacy of the intervention among subgroups of study subjects, follow-up rates were stratified according to baseline demographic and other variables (Table 3). Overall follow-up rates did not differ significantly by gender, race, age, insurance, or primary care type. Among the remaining independent variables, multiple previous asthma ED visits, hospitalizations, and the use of daily preventive medications at baseline did correlate with an increased likelihood of follow-up. Stratified relative probabilities for follow-up did not differ significantly when compared by Mantel-Haenszel test for homogeneity. For adjusting for differences in baseline variables that may have been associated with follow-up, a multiple logistic regression including the variables listed in Table 3 was performed. The adjusted odds ratio for the intervention (1.8; 95% CI: 1.1–3.0) was similar to the unadjusted odds ratio (2.0; 95% CI: 1.3–3.3).
To assess the effect of the intervention on other outcomes, we followed subsequent PCP visits, ED visits, and hospitalizations for 1 year among a subgroup of 135 subjects (67 control and 68 intervention) who were followed by primary care sites within our hospital system. This subgroup was similar to the overall study group in demographic composition and asthma history; the proportion of subjects with Medical Assistance or no insurance was higher for this subgroup than for other subjects (76% vs 62%; P = .008). There was a trend toward a greater number of PCP visits during the subsequent year among intervention subjects, although this did not reach the level of statistical significance (control median 1, interquartile range [0,2] vs intervention 1 [0,3] P = .08). Similar numbers of intervention (36) and control (32) subjects had an asthma-related ED visit during the subsequent year (53% vs 48%; P = .5). A greater number of intervention subjects (21) compared with controls (10) were hospitalized for asthma during this year, and the difference reached the level of statistical significance (31% vs 15%; P = .03).
Overall, the conclusions of this study supplement previous work that has documented poor linkage between ED and primary care for urban children with asthma. First, our results demonstrate a low baseline rate of follow-up similar to that observed in previous studies. Second, providing a follow-up appointment increased the rate of follow-up, consistent with the hypothesis that gaining access to an established source of primary care is a significant barrier for urban families. Finally, although follow-up rates improved, we found little improvement in short-term outcomes or use of controller medications. This result suggests that additional research is needed to develop additional interventions in this population.
The low PCP follow-up rate measured among urban children here is similar to what has been observed in previous descriptive studies. The most comprehensive study to examine this issue was the National Cooperative Inner-City Asthma Study, which enrolled a total of 1528 children with asthma between the ages of 4 and 9 from 8 inner-city areas.13,14 In-depth interviews were conducted with the parents of 344 of these children 3 to 5 weeks after an ED asthma visit. On the basis of parental report, the investigators estimated that 52% of the children were seen by a PCP during the month after the ED visit, similar to the 51% follow-up rate reported by parents of controls in our study. Our results further suggest that this may be an overestimate, as the PCP follow-up rate fell to 44% among control subjects when parental report was verified with the provider. We believed that verification of parental report was important as parents may be reluctant to admit nonadherence to physician recommendations. Future research in this area may benefit from similar verification, just as the validity of studies of asthma medication adherence has been improved with the use of automatic dose-counting devices.15
Another benefit of verification was the ability to determine the actual date of the next PCP visit. The shape of the curves in Fig 2 demonstrates the importance of the weeks after the ED visit as a window of opportunity to link with primary care. Most children who were not seen during the month after the visit also did not see a provider during the subsequent 5 months. Many of these later PCP visits were for well-child care when asthma was 1 of many topics to be discussed. The most common reason given for lack of PCP follow-up in both the National Cooperative and our study was that the child had improved and a follow-up visit was not needed. Although the acute exacerbation may indeed have improved, this answer suggests a lack of understanding of the chronic nature of asthma and the benefits of preventive care. Focus groups with inner-city parents suggest that beliefs and misconceptions about asthma are among the most important barriers to establishing ongoing care.16 Although changing these misconceptions may require comprehensive educational interventions, the ED visit may be a place to begin these efforts. The heightened parental concern leading to an ED visit may present a “teachable moment,” a chance to focus the family’s attention on improving future care of the child’s asthma.
Few previous studies have examined ED-based interventions to improve follow-up for asthma. Zeiger et al17 randomly assigned pediatric and adult ED asthma patients to an allergy referral and observed improved long-term outcomes. However, the transition to Medicaid managed care (in Pennsylvania and other states) and an accompanying focus on primary care–based treatment of asthma would not allow this to be a feasible option in the ED setting. Baren et al18 recently assessed a 3-part intervention for adult ED asthma patients consisting of a course of oral prednisone, a reminder telephone call, and a voucher for transportation to the follow-up appointment. Intervention subjects were more likely to follow-up with a PCP compared with controls (46% vs 29%). The authors were unable to determine which part of the intervention was most effective. Although this intervention was tailored to adults, children may have different barriers to overcome. In the National Cooperative Inner-City Asthma Study, inner-city parents more commonly reported health system and organizational barriers such as lack of available appointments rather than transportation problems or inability to pay for care.14 Therefore, we tailored our intervention to ensuring that a PCP appointment had been scheduled.
The results of our trial are both encouraging and cause for concern. Providing a follow-up appointment after an ED visit does seem to improve the likelihood that a child will see a PCP. However, similar to what parents had reported in previous studies, the study staff encountered considerable difficulty in helping to arrange these appointments. In the majority (76%) of cases, an appointment could not be obtained at the time of the ED visit, particularly when it occurred after regular office hours. Once the patient had left the ED, multiple telephone calls were often required to ensure that an appointment had been scheduled. The difficulty encountered in completing the intervention would complicate implementation into routine practice. Improvements in communication systems and access to office schedules from the ED would be required for this to be a feasible routine outside a research protocol.
Targeting the intervention to a specific group of patients most at risk for lack of follow-up would be another option. Although the efficacy of the intervention did not differ significantly across patient subgroups with this sample size, some trends did emerge. Control subjects with Medical Assistance had lower follow-up rates compared with those with commercial insurance (41% vs 58%), and the impact of the intervention seemed to be greater (relative risk: 1.5 vs 1.1). Overall, subjects with recurrent ED visits or a hospitalization were more likely to follow up with their PCP; this may simply reflect greater asthma severity and increased parental concern. Children who were already using a daily preventive medication were also more likely to make a follow-up visit. This may also reflect increased severity or, alternatively, an already established continuity relationship with their PCP. In no subgroup analyzed (except for the 17 nonblack subjects) did >60% of control subjects make a follow-up appointment. This low baseline follow-up rate suggests that a broad-based intervention across this population would be most appropriate.
Providing a follow-up appointment did not seem to increase the overall use of preventive controller medications. We chose to measure this as a study outcome because previous studies have demonstrated the efficacy of these medications, particularly inhaled corticosteroids, in reducing asthma morbidity.7–9 The NAEPP Guidelines recommend a controller medication for children with persistent symptoms, defined as symptoms that occur more than twice per week during the day or twice per month during the night. At the time of the ED visit, families described persistent symptoms in the majority (70%) of children enrolled in the study; however, only 38% reported use of a daily preventive medication. Four weeks after the ED visit, the rate had increased to only 56% overall with no significant difference between the intervention subjects and controls. There was a trend toward increased use when the analysis was limited to subjects who had described persistent asthma symptoms at the ED visit (64% vs 53%; P = .11); this might have reached the level of statistical significance in a larger study. Our ability to assess the appropriateness of prescribed therapy at follow-up was limited. The degree of symptoms as well as parental perceptions of symptoms may have changed over the weeks between the acute visit and follow-up. In addition, although the NAEPP Guidelines focus on chronic symptoms, an exacerbation resulting in an ED visit could be viewed as an appropriate indication for beginning controller therapy. Although not conclusive, our results suggest several areas for future work. First, consistent with previous studies, preventive medications are underused in this population.10–12 Second, opportunities to prescribe preventive medications are being missed at the time of follow-up. More than one third of the children who had persistent symptoms and did report a follow-up visit did not begin a preventive medication after that visit. More comprehensive interventions involving provider education may be needed to reduce these missed opportunities. Finally, controller medication use remained low among subjects who had persistent asthma and did not make a follow-up visit (46%). Additional ED-based studies to identify these patients and initiate preventive therapies have the potential to reduce future morbidity. In particular, initiation of controller medications in the ED may be appropriate in some cases and should be studied further.
In a subgroup analysis, the study intervention did not seem to improve outcomes as measured by ED visits during the year after the initial ED visit. Intervention subjects were actually more likely to be hospitalized (31% vs 15%), although this may have been attributable to chance as the percentage in each group returning for an ED visit was very similar (53% vs 48%). Future studies should measure these traditional outcomes as well as others such as quality of life and asthma-related disability, which are of great importance to families.
There were a number of limitations to this study. First, patients were enrolled only if they were being discharged between the hours of 8:00 am and midnight. It is possible that patients who seek asthma treatment during late evening hours may differ from the patients enrolled in this study. Second, as discussed above, it was not possible to obtain appointments during the ED visit for a majority of patients. It is possible that the repeated telephone calls made to families in the intervention group may have provided an additional incentive to follow-up with a PCP beyond what would have been present if a simple appointment had been provided. Replication of this study at another or multiple institutions where appointments could be more readily obtained might serve to address this issue as well as increase the generalizability of the results.
Overall, the results of this study demonstrate that providing a follow-up primary care appointment after an ED visit is associated with an increased likelihood of follow-up. Although follow-up increased in the intervention group, this increase was not associated with improved short-term outcomes or the use of preventive medications. Future research should focus on improving links between ED and primary care to make this intervention more feasible as well as assessing other ED-based interventions that may serve to improve ongoing care for childhood asthma.
This study was funded by the Pew Charitable Trusts and the University of Pennsylvania Research Foundation
- ↵National Asthma Education and Prevention Program. Expert Panel Report 2: Guidelines for the Diagnosis and Management of Asthma. Bethesda, MD: National Institutes of Health; 1997. Publ. No. 98-4051
- ↵Lieu TA, Quesenberry CP, Capra AM, Sorel ME, Martin KE, Mendoza GR. Outpatient management practices associated with reduced risk of pediatric asthma hospitalization and emergency department visits. Pediatrics.1997;100 :334– 341
- ↵Adams RJ, Fuhlbrigge A, Finkelstein JA, et al. Impact of inhaled anti-inflammatory therapy on hospitalization and emergency department visits for children with asthma. Pediatrics.2001;107 :706– 711
- ↵Warman KL, Johnson Silver E, McCourt MP, Stein REK. How does home management of asthma exacerbations by parents of inner-city children differ from NHLBI Guideline Recommendations? Pediatrics.1999;103 :422– 427
- ↵Scarfone RJ, Zorc JJ, Capraro GA. Patient self-management of acute asthma: adherence to national guidelines a decade later. Pediatrics.2001;108 :1332– 1338
- ↵Leickly FE, Wade SL, Crain E, Kruszon-Moran D, Wright EC, Evans R. Self-reported adherence, management behavior, and barriers to care after an emergency department visit by inner city children with asthma. Pediatrics.1998;101(5) . Available at: www.pediatrics.org/cgi/content/full/101/5/e8
- ↵Mansour ME, Lanphear BP, DeWitt TG. Barriers to asthma care in urban children: parent perspectives. Pediatrics.2000;106 :512– 519
- Copyright © 2003 by the American Academy of Pediatrics