A randomized controlled trial of a pediatric asthma outreach program

A randomized controlled trial of a pediatric asthma outreach program Dirk K. Greineder, MD, PhD, a,b Kathleen C. Loane, RN, c and Paula Parks, RN, NP a Boston, Mass Background: Previous studies have shown that asthma education and case management may reduce asthma emergency care, hospitalizations, and expenditures. Objective: We sought to study the effect of an asthma outreach program (AOP), a team-based, case-management intervention, on emergency ward (EW) and hospital use. Methods: Fifty-seven patients aged 1 to 15 years with the diagnosis of asthma based on the usual clinical practice criteria who were continuously enrolled in a staff-model health maintenance organization for a period of at least 2 consecutive years were randomized into 2 intervention groups. The control group received a single intensive asthma education intervention, and the AOP group received the same initial education but then was followed-up by an asthma case management nurse throughout the intervention period. Results: EW visits, hospitalizations, and total outside-ofhealth-plan expenditures (consisting of EW and hospital expenses, as well as miscellaneous costs, such as ambulance, durable medical equipment, tertiary referrals, and home care) were assessed from claims filed for a year before and after enrollment. Control group patients experienced significant reductions in EW visits (39%), hospitalizations (43%), and outside-of-health-plan costs (28%), possibly as a result of the baseline educational intervention received by all enrolled patients, in conjunction with regression to the mean. AOP group patients experienced significant reductions in EW visits, (73%, P =.0002), hospitalizations (84%, P =.0012), and outside-of-health-plan use (82%, P <.0001). When compared with the control group, AOP group patients demonstrated additional significant reductions in EW visits (57%, P <.05), hospitalizations (75%, P <.05), and outside-of-health-plan use (71%, P <.001). Estimates of direct savings to the health plan ranged from $7.69 to $11.67 for every dollar spent on the AOP nurse s salary, depending on assumptions. Conclusions: Asthma patients in a staff-model health maintenance organization decreased their resource use between 57% to 75% by participation in an AOP as compared with a randomized control group receiving only an educational intervention. Substantial savings were achieved compared with the cost of the AOP nurse. (J Allergy Clin Immunol 1999;103:436-40.) Key words: Asthma, child, case management, health education, cost, risk, managed care, allergy, outreach From a the Department of Allergy, Harvard Pilgrim Health Care, Boston; b the Division of Rheumatology, Immunology and Allergy, Department of Medicine, Brigham and Women s Hospital, Boston; and c the Department of Ambulatory Care and Prevention, Harvard Medical School, Boston. Supported in part by the Harvard Pilgrim Health Care Foundation. Received for publication Aug 14, 1998; revised Oct 19, 1998; accepted for publication Oct 22, 1998. Reprint requests: Dirk K. Greineder, MD, PhD, Department of Allergy, Harvard Pilgrim Health Care, 133 Brookline Ave, Boston, MA 02215. Copyright 1999 by Mosby, Inc. 0091-6749/99 $8.00 + 0 1/1/96025 436 Abbreviations used AOP: Asthma outreach program CI: Confidence interval EW: Emergency ward Asthma constitutes a major pediatric public health problem in the United States. 1-7 It is the most common chronic disease of childhood, affecting an estimated 4.8 million children, 8 and in 1990, it caused the estimated loss of 10 million school days and resulted in total pediatric asthma costs approximated at 1.8 billion dollars. 7 More than 5000 people die of asthma annually, and asthma hospitalization rates are highest among children and black subjects. 9 We have previously reported on the reduction of hospital and emergency ward (EW) resource use by an asthma outreach program (AOP). 10 In this report we wish to provide data regarding a similar cohort of patients who were studied in a randomized controlled trial. In our original report we were only able to analyze resource use in a before and after trial design because all patients referred to the AOP were enrolled at the onset of the intervention. In this study patients referred to the AOP were randomized into 2 groups; only one of the groups was provided the full AOP intervention, whereas the control group was given a single asthma education class by the AOP nurse. As will be apparent, even the control group had a reduction in health care use, possibly as a result of the class and their identification as a highrisk group requiring referral. However, the reduction in health care use by the AOP group was much greater. In addition, in this study we have been able to obtain outside health care use expense reports (consisting of total out-of-health-plan costs, including EW and hospital expenses, as well as miscellaneous costs, such as ambulance, durable medical equipment, tertiary referrals, and home care). These reports summarize the economic impact of the AOP in addition to the quality of care benefits that derive from protecting children from the risks associated with EW and hospital visits for acute asthma. METHODS A total of 57 pediatric patients with asthma between the age of 1 and 15 years were enrolled in this phase of the AOP. Patients less than 12 months of age at randomization who had wheezing for the first time or who had the eventual diagnosis of bronchopulmonary dysplasia were excluded. Patients were selected from predominately urban health centers of the Health Centers Division of Harvard Pilgrim Health Care, a large health maintenance organization in New England. This study was approved by the Investigational

J ALLERGY CLIN IMMUNOL VOLUME 103, NUMBER 3, PART 1 Greineder, Loane, and Parks 437 Review Board of Harvard Pilgrim Health Care. Patients were recruited by reviewing hospitalization lists for patients admitted with the diagnosis of acute asthma, with or without status asthmaticus (group H). Patients were also referred by their primary care pediatricians on the basis of subjective impressions that these patients were at risk for asthma exacerbations (group R). Both groups were independently randomized by using a paired randomization scheme based on a table of random numbers such that equal numbers of each type of patient would be assured in the control and AOP groups. After randomization, both groups received an educational intervention; however, only the AOP group received longterm follow-up. For this report, patients were enrolled in the AOP during a 12- month period and followed-up for an additional year, assuring at least 1 year of follow-up after enrollment for each patient. Consequently, the after period was between 1 to 2 years long, depending on the initial time of enrollment. No enrolled patients terminated their coverage with the Health Centers Division during the study period, and all were members at least 1 year before enrollment. Before and after observation periods for each patient were of equal length (at least 1 year). The AOP consists of an experienced allergy nurse, an allergy nurse practitioner, and an allergist who is available for guidance and consultation and has been described in our previous report. 10 On referral, the patient s family is called by the asthma outreach nurse and then scheduled for a visit with her. If patients fail to keep this initial appointment, the asthma nurse continues to call the family until the visit is successfully kept. The initial visit is customized to the individual family s needs to assure that at the conclusion there is a common knowledge base. The patient and family are given oneon-one education in 7 domains: asthma definition and pathogenesis; triggers and warning signs; asthma medications; inhaler, spacer, and nebulizer training as indicated; use of peak flow meter and zoning instruction; environmental control, including smoking and allergens; and adherence with medication schedules and physician visits. Teaching is by discussion, direct demonstration, and backdemonstration and is supported by the use of commonly available handouts as deemed appropriate by the outreach nurse. Inhaler and peak flow meter technique are reviewed, and zoning instructions (based on National Asthma Education Program guidelines 11 ) are delivered if the child is able to perform reliable peak flow maneuvers. A written asthma control plan based on the patient s current medications is provided and reviewed with the family. For this study, the AOP nurse applied the National Heart, Lung, and Blood Institute Guidelines for the Diagnosis and Management of Asthma 11 (the 1997 revision was not yet available) when assessing the adequacy of the asthma control plan and, if the medication plan was felt to be inadequate or not conforming, obtained appropriate orders from the pediatrician to ensure compliance with the guidelines. Both groups (control and AOP) had equal access to medications because most health plan members have prepaid prescription plans. Where appropriate, environmental allergen and smoking avoidance instructions are provided. Both groups were equally referred for allergy consultation when allergy was likely to contribute to the asthma severity on the basis of history or when the patient lived in a high-risk environment for allergic complications (eg, with pets or roaches or in a basement apartment). This initial visit typically lasts several hours. Both control patients and AOP patients were educated in this manner because the Human Subjects Committee felt that our preliminary study 10 mandated an intervention for all participants in the study, even though at that time such a resource was not available to nonstudy patients. Finally, independent referral to the allergy or pulmonary departments were equally available to both groups. The main difference between the control group and the AOP group is the follow-up. The control group is referred back to FIG 1. Comparison of the mean use ± SEM of EW visits, hospitalizations, and outside-plan dollar expenses between the control and AOP group after enrollment. These differences represent 57%, 75%, and 71% reductions in use and are significant (P <.05, P <.05, and P <.001, respectively). their pediatrician for ongoing care after the initial educational visit and is not further contacted by the AOP nurse. By contrast, the nurse remains in close contact with the AOP group. Much of the followup is done on the telephone, although occasional additional visits are scheduled as deemed necessary by the nurse. This follow-up assures that patients are keeping routine pediatric appointments, taking maintenance medication, and keeping peak flow charts as requested. Action plans are updated as needed. In addition, the nurse remains a resource and counselor for the family. Further details of the outreach intervention have been previously published. 10 For this report, in addition to the EW visit and hospitalization data, we were able to obtain actual outside-of-health-plan use costs (ie, costs not part of the staff model health maintenance organization s internal costs). These costs were captured by extracting data related to a wide range of asthma-related diagnoses from the claims computer for each patient enrolled in the study. Data were collected starting 2 years before the first patient was enrolled and continuing for 1 year after the last patient was enrolled to assure that all data was available for an equal time before and after enrollment for each patient. Data extraction was refined until only asthma-related illness claims were included. This reflects money spent on hospital costs and EW visits (but not internal, health center-based, urgent care or routine ambulatory care), as well as outside referrals, home care, and durable medical equipment, such as nebulizers and supplies. This allows direct outside-plan cost comparisons to be made between the different groups and interventions. Numbers of hospitalizations and EW visits before and after enrollment were counted manually from a printout of the data from the claims computer. No effort was made to track the internal, health center based use patterns of the 2 groups before or after enrollment because it was not our intention to alter or study routine care costs. Statistical analysis was performed with the Wilcoxon matchedpair signed-rank test 12 (for comparisons within groups before and after intervention) and by the chi-squared test (for comparisons between groups). 13

438 Greineder, Loane, and Parks J ALLERGY CLIN IMMUNOL MARCH 1999 TABLE I. Resource use in control patients EW visits Hosp OU dollars Patient no Sex Age* Race Group Before After Before After Before After 1 F 3 Asian H 3 5 4 3 $9638 $12,251 2 M 2 Black H 0 0 1 0 $841 $0 3 M 4 Hispanic H 1 1 1 2 $1969 $2114 4 F 1 Black H 1 0 1 0 $878 $80 5 F 4 Black H 1 0 1 0 $991 $0 6 F 1 Black H 1 1 1 1 $2330 $1103 7 F 4 Hispanic H 3 1 3 1 $5800 $1036 8 F 12 Black H 1 0 1 0 $2264 $0 9 M 1 White H 1 0 1 0 $3285 $0 10 F 8 Black R 1 0 0 0 $239 $0 11 M 4 White R 1 0 0 0 $302 $0 12 M 12 White R 1 0 0 0 $380 $0 13 F 9 Black R 3 0 1 0 $3278 $0 14 M 6 Hispanic R 4 5 4 5 $9542 $20,569 15 F 1 Black R 3 1 2 0 $3682 $175 16 F 14 Black R 1 1 0 0 $214 $197 17 M 11 White R 3 1 1 0 $1788 $91 18 M 7 Black R 0 1 1 0 $2095 $62 19 M 11 White R 4 4 2 2 $2851 $3046 20 M 1 White R 2 0 1 0 $5042 $0 21 M 1 Hispanic R 1 1 1 1 $2077 $3021 22 M 1 Hispanic R 2 1 0 1 $506 $1272 23 M 1 Black R 1 0 0 0 $196 $0 24 F 4 Black R 2 0 0 0 $846 $105 25 M 2 White R 2 1 1 0 $2073 $292 26 F 4 Asian R 1 3 0 0 $343 $448 27 F 11 Black R 0 0 0 0 $0 $0 28 M 2 Hispanic R 0 0 0 0 $0 $0 Sum 44 27 28 16 $63,450 $45,862 % Change (after) 39 43 28 Mean 1.57 0.96 1.00 0.57 $2266 $1638 Standard error 0.22 0.28 0.21 0.22 $484 $834 P value.025.019.003 Hosp, Overnight hospitalizations; OU, outside-plan use; H, patients randomized to control intervention after a hospitalization; R, patients randomized to a control intervention after referral from pediatrician. *Age is in years, rounded to the nearest integer, at the time of enrollment into the AOP. Determined by signed-rank test comparing before/after. RESULTS Fifty-seven patients were enrolled in this randomized trial as defined above. A total of 18 patients were recruited from the hospitalization list, with 9 being randomized into the control group (group H, Table I) and an equal number into the AOP group (group H, Table II). An additional 39 patients were obtained by referral from pediatricians and were randomized independently into the control (group R, Table I) and AOP (group R, Table II) groups. As expected, the patients derived from the hospitalization lists had slightly more severe asthma than the remaining patients, although there are no dramatic differences between the groups. The overall results are summarized in the Tables. Note that mean EW use, hospitalizations, and cost per patient before intervention are similar for the control (Table I) and AOP (Table II) groups. EW use is reduced by 39% (P =.025) after enrollment in the control group, with 44 visits before intervention (mean = 1.57; 95% confidence interval [CI] = 1.12-2.02) and 27 visits after intervention (mean = 0.96; 95% CI = 0.39-1.53). In the AOP group EW use was reduced by 73% (P =.0002) after enrollment, with 45 visits before intervention (mean = 1.55; 95% CI = 1.00-2.10) and 12 visits after intervention (mean = 0.41; 95% CI = 0.19-0.63). Similarly, hospitalization was reduced by 43% (P =.019) after enrollment in the control group, with 28 admissions before intervention (mean = 1.00; 95% CI = 0.57-1.43) and 16 admissions after intervention (mean = 0.57; 95% CI = 0.12-1.02). In the AOP group hospitalization was reduced by 84% (P =.0012) after enrollment, with 25 admissions before intervention (mean = 0.86; 95% CI = 0.50-1.22) and 4 admissions after intervention (mean = 0.14; 95% CI = 0.03-0.31). Total outside-plan use in dollars was reduced by 28% (P =.003) after enrollment in the control group, with $63,450 expended before intervention (mean = $2266; 95% CI = $1317- $3214) and $45,862 after intervention (mean = $1638; 95% CI = $3-$3273). In the AOP group outside-plan use was reduced by 82% (P <.0001) after enrollment, with $78,070 spent before intervention (mean = $2692; 95%

J ALLERGY CLIN IMMUNOL VOLUME 103, NUMBER 3, PART 1 Greineder, Loane, and Parks 439 TABLE II. Resource use in AOP patients EW visits Hosp OU dollars Patient no Sex Age* Race Group Before After Before After Before After 1 M 4 Black H 5 1 3 0 $3681 $313 2 M 3 White H 2 0 1 0 $3418 $0 3 F 12 Black H 2 1 1 0 $3187 $306 4 M 2 White H 5 0 3 0 $5208 $80 5 M 5 Black H 1 1 1 0 $2771 $229 6 F 1 Black H 2 0 1 0 $1505 $168 7 M 1 Black H 0 0 1 0 $8042 $0 8 F 4 Black H 2 1 1 0 $2467 $126 9 F 1 Black H 3 2 2 2 $10,679 $6790 10 M 2 Black R 3 0 1 0 $2577 $0 11 M 3 White R 1 1 1 1 $3721 $1993 12 M 4 Black R 1 0 0 0 $263 $0 13 F 2 Black R 2 0 1 0 $2676 $0 14 M 1 Hispanic R 2 0 1 0 $1929 $0 15 M 14 Black R 2 1 1 1 $3805 $1994 16 F 15 White R 1 1 2 0 $8181 $249 17 F 1 White R 2 1 0 0 $782 $251 18 M 1 Black R 4 1 3 0 $10,907 $578 19 M 1 Black R 1 0 0 0 $317 $0 20 F 8 Black R 2 0 0 0 $734 $52 21 F 1 Black R 0 1 0 0 $131 $543 22 M 1 White R 2 0 1 0 $1089 $0 23 M 5 Black R 0 0 0 0 $0 $0 24 M 8 Black R 0 0 0 0 $0 $0 25 M 7 White R 0 0 0 0 $0 $0 26 F 4 White R 0 0 0 0 $0 $0 27 F 2 Black R 0 0 0 0 $0 $0 28 F 1 Black R 0 0 0 0 $0 $0 29 M 1 Black R 0 0 0 0 $0 $0 Sum 45 12 25 4 $78,070 $13,672 % Change (after) 73 84 82 Mean 1.55 0.41 0.86 0.14 $2692 $471 Standard error 0.27 0.11 0.18 0.08 $589 $245 P value.0002.0012 <.0001 Hosp, Overnight hospitalizations; OU, outside-plan use; H, patients randomized to AOP intervention after a hospitalization; R, patients randomized to AOP intervention after referral from a pediatrician. *Age is in years, rounded to the nearest integer, at the time of enrollment into the AOP. Determined by signed-rank test comparing before/after. CI = $1538-$3846) and $13,672 after intervention (mean = $471; 95% CI = $ 9-$951). The results comparing the mean resource use of the control and AOP groups during the observation period after intervention are graphically summarized in Fig 1. Compared with the control group, the AOP group use for EW visits, hospitalizations, and outside expenditures is reduced by 57%, 75%, and 71%, respectively. Chi-squared analysis comparing use before and after enrollment in the 2 groups reveals that the reductions in EW visits, hospitalizations, and outside-plan use in the AOP group are all significantly different from the control group (P <.05, P <.05, and P <.001, respectively). The asthma outreach nurse managing the patients reported in this trial was also following approximately 99 other patients for asthma case management while working 0.40 of a full-time equivalent. At an annual salary at the time of the study of approximately $50,000, with fringe benefit costs of 20%, this represents a total cost of $24,000. The 29 AOP patients from this trial represent 23% of the total 128 patients actively followed-up by the asthma outreach nurse during this time (not counting the additional 28 control patients), suggesting that the additional case management labor cost for these patients totaled approximately $5520. The difference in outsideplan costs for the 29 AOP patients before and after intervention was $64,398. Therefore $11.67 is saved for every dollar spent on the asthma outreach nurse. This value is similar to the one seen in our original report. 10 DISCUSSION The before and after intervention results for the AOP group are similar to those obtained in our previously reported uncontrolled trial. 10 In that study we did not have a control group and therefore had no effective way to estimate the influence of natural regression to the mean or what effect the result of identifying patients as

440 Greineder, Loane, and Parks J ALLERGY CLIN IMMUNOL MARCH 1999 high risk (by their pediatrician or because of the patients EW use or hospitalizations) would have on subsequent use. We have independent observations on a similar population in our health plan that suggest that at least some of the reduction in use may be the result of such influences. 14 The control group in this study does indeed manifest a statistically significant reduction in all aspects of use, ranging between 28% and 43%. We assume that this result derives from a combination of the factors discussed above and the very intense educational session that these patients and families attended on enrollment into the control group. At least 1 previous study has shown that a single educational intervention can have substantial effects in improving families asthma knowledge and behaviors, although in that study clinical outcomes were not improved. 15 The control group effect was significantly enhanced by the full-scale intervention of the AOP, which produced significantly better gains than the control intervention. Others have reported on educational and home care activities to reduce health care use, with variable results. Several studies with ongoing interventions reminiscent of our outreach program improved asthma knowledge and reduced use, 16,17 whereas others showed improved knowledge but had no impact on use, 15,18 and at least 1 study showed improved knowledge but increased use. 19 We calculated an estimate of $11.67 saved for every dollar spent on the asthma nurse (see the Results section). However, on the basis of the 28% reduction of costs in the control group, it is possible that 34% (28 of 82) of these savings can be attributed to the control intervention coupled with the spontaneous decrease associated with regression to the mean and/or the identification of a high-risk patient. Even with that correction, $7.69 is saved for every dollar spent on the asthma outreach nurse. No additional correction was made for the time commitment by the allergist and nurse practitioner because the support required after the initial AOP nurse training was slight and typically provided informally, with no reduction in productivity for these clinicians. Additionally, we have not calculated potentially increased costs associated with possibly increased frequencies of routine office visits and better medication adherence by AOP patients. These are costs associated with appropriate care as defined by National Heart, Lung, and Blood Institute guidelines. 11 In addition to the decrease in costs, it is likely that quality of care and quality of life, although not formally measured, are improved by reducing the frequency of EW visits or hospital admissions. Each of these events represents a high-risk and high-stress situation for the patient, the family, and the health care delivery system. Finally, although not specifically studied, feedback clearly indicates that patient and family satisfaction with care is markedly enhanced by the AOP. We thank all the pediatricians who referred patients to the project and cared for those who were randomized to the control group. REFERENCES 1. Gergen PJ, Weiss KB. Changing patterns of asthma hospitalization among children: 1979 to 1987. JAMA 1990;264:1688-92. 2. Taylor WR, Newacheck PW. Impact of childhood asthma on health. Pediatrics 1992;90:657-62. 3. Evans RE, Mullally DI, Wilson RW, et al. National trends in the morbidity and mortality of asthma over two decades: 1965-1984. Chest 1987;91:65S-74S. 4. Gergen PJ, Mullally DI, Evans R. National survey of prevalence of asthma among children in the United States, 1976 to 1980. Pediatrics 1988;81:1-7. 5. Halfon N, Newacheck PW. Trends in the hospitalization for acute childhood asthma, 1970-84. Am J Public Health. 1986;76:1308-11. 6. Weiss KB, Zimmerman EM, Gergen PJ, et al. Special report of the NIAID task force on immunology and allergy: the socioeconomic impact of immunologic and allergic diseases in the United States. Washington (DC): US Dept of Health and Human Services; 1991. 7. Weiss KB, Gergen PJ, Hodgson TA. An economic evaluation of asthma in the United States. N Engl J Med 1992;326:862-6. 8. Centers for Disease Control and Prevention. Asthma-United States, 1989-1992. MMWR 1995;43:952-5. 9. Centers for Disease Control and Prevention. Asthma mortality and hospitalization among children and young adults-united States, 1990-1993. MMWR 1996;45:350-3. 10. Greineder DK, Loane KC, Parks P. Reduction in resource utilization by an asthma outreach program. Arch Pediatr Adolesc Med 1995;149:415-20. 11. National Asthma Education Program. Guidelines for the Diagnosis and Management of Asthma. Bethesda (Md): US Dept of Health and Human Services; 1991. NIH publication no. 91-3642. 12. Siegel S. Nonparametric statistics for the behavioral sciences. New York: McGraw-Hill Book Co; 1956. 13. Swinscow TDV. Statistics at square one. London: British Medical Association; 1978. 14. Greineder DK, Loane KC, Parks P. Outcomes for control patients referred to a pediatric asthma outreach program: an example of the Hawthorne Effect. Am J Man Care 1998;4:196-202. 15. Garrett J, Fenwick JM, Taylor G, Mitchell E, Stewart J, Rea H. Prospective controlled evaluation of the effect of a community based asthma education centre in a multiracial working class environment. Thorax 1994;49:976-83. 16. Fireman P, Friday GA, Gira C, Vierthaler WA, Michaels L. Teaching selfmanagement skills to asthmatic children and their parents in an ambulatory care setting. Pediatrics 1981;68:341-8. 17. Hughes DM, McLoed M, Garner B, Goldbloom RB. Controlled trial of a home and ambulatory program for asthmatic children. Pediatrics 1991;87:54-61. 18. Clark NM, Feldman CH, Evans D, Levison MJ, Wasilewski Y, Mellins RB. The impact of health education on frequency and cost of health care use by low income children with asthma. J Allergy Clin Immunol 1986;78:108-15. 19. Mitchell EA, Britton JR, Tattersfield AE. Management of asthma in schools. Arch Dis Child 1987;62:414-5.