Race and Sex Differences in Consistency of Care With National Asthma Guidelines in Managed Care Organizations

Background  In the United States, morbidity from asthma disproportionately affects African Americans and women. Although inadequate care contributes to overall asthma morbidity, less is known about differences in asthma care by race and sex.

Subjects and Methods  To examine the relationships of race and sex with asthma care, we analyzed responses to questionnaires administered to adults enrolled in 16 managed care organizations participating in the Outcomes Management System Asthma Study between September and December 1993. Indicators of care consistent with National Asthma Education and Prevention Program (1991) recommendations were assessed. Of a random sample of 8640 patients asked to participate, 6612 (77%) completed the survey. This study focused on 5062 (14% African American, 72% women) patients with at least moderate asthma symptom severity.

Results  Fewer African Americans than whites reported care consistent with recommendations for medication use (eg, daily inhaled corticosteroid use, 34.9% vs 54.4%; P = .001), self-management education (eg, action plan, 42.0% vs 53.8%; P = .001), avoiding triggers (37.6% vs 53.6%; P = .001), and specialist care (28.3% vs 41.0%; P = .001). Differences in asthma care by sex were smaller and tended to favor women except for daily inhaled corticosteroid use (women vs men: 49.6% vs 58.3%; P = .001) and having specialist care (37.7% vs 43.1%; P = .001). Similar race and sex differences were observed after adjusting for age, education, employment, and symptom frequency.

Conclusions  Even among patients with health insurance, disparities in asthma care for African Americans compared with whites exist and may contribute to race disparities in outcomes. Women generally reported better asthma care but may benefit from greater use of inhaled corticosteroids.

ASTHMA, A CHRONIC disease characterized by airway inflammation, affects 14 to 15 million people in the United States,¹ and accounts annually for 1.2 million emergency department (ED) visits, 445 000 hospital days, and an economic burden of $5.1 billion.² Although there are effective therapies for asthma, inadequate symptom control remains a problem for many patients,³ particularly for African Americans^4-8 and women.^8-10 For example, asthma-related hospitalization and mortality in African Americans are 1.4 to 4.0 times^6,11,12 and 1.3 to 5.5 times^4,5,11,13,14 more likely, respectively, than in whites. Women with asthma report significantly lower quality of life¹⁰ than men and have 2.5 to 3.0 times the rate of hospitalization.⁹ Studies of other chronic diseases suggest that differences in medical care may contribute to variations in outcomes by race and sex.^15-19 While inadequate or inappropriate therapy is known to contribute to morbidity and mortality from asthma in the United States,³ the relationships of race and sex to asthma care are incompletely understood.

Some studies have suggested that poor outcomes among African Americans with asthma may reflect socioeconomic factors,^5,13,20 including financial barriers to adequate care. However, others have found that differences in socioeconomic status and health insurance coverage between patients only partially explain race differences in health care.^6,12,21,22 Two recent studies^8,23 in patients with asthma reported that African Americans enrolled in managed care organizations (MCOs) were less likely to use inhaled corticosteroids (ICS) than whites, suggesting that ineffective patterns of medication use may contribute to race disparities in asthma outcomes even among patients with access to health care. National guidelines for asthma care (1991 National Asthma Education and Prevention Program [NAEPP],³ revised in 1997²⁴), however, emphasize that appropriate medication use is only one aspect of an effective strategy to improve clinical outcomes. These guidelines highlight the importance of encouraging active patient participation in a partnership (patient-physician partnership) for care, including patient education for self-management, control of triggers, and periodic assessment of symptom control to reduce the frequency of asthma exacerbations,^25-27 ED visits,²⁸ hospitalizations,^28-31 and missed workdays.³² Whether these aspects of asthma care vary by race has not been examined.

The basis for greater morbidity in women compared with men with asthma is also unexplained. Some evidence suggests that dissimilar environmental exposures and host susceptibility factors may contribute to sex differences in asthma outcomes.³³ A recent study found that women were less likely than men to use multidose inhalers properly,³⁴ suggesting that fewer women are benefiting from targeted delivery of asthma medications. As with race, however, it is not known whether there are other important sex-related differences in the comprehensive approach to asthma care recommended by the guidelines.

The purpose of this study was to examine the relationships of race and sex to a comprehensive array of guideline recommendations for asthma care. To disentangle the effects of health insurance on asthma care from those due to race and sex, we assessed these relationships in adults enrolled in employer-based MCOs.

This was a cross-sectional study using patient-reported survey data from the Managed Health Care Association Outcomes Management System Asthma Study (MHCA study).^35,36 The MHCA study was undertaken by 11 large US corporations and their managed care partners to test the feasibility and usefulness of patient-reported information to identify opportunities to improve the quality of asthma care.

The survey instrument was constructed by the MHCA Asthma Study group based largely on the Asthma TyPE instrument developed by the Health Outcomes Institute (1994). Additional items were added, including the impact of asthma on daily life, self-management knowledge, and ratings of access to care. In a feasibility study conducted in 962 patients,³⁷ there was moderately high concordance between patient and physician reports of medication use (80.1% for inhaled β₂-agonists, 73.4% for oral methylxanthines, and 81.7% for ICS). In the MHCA study (source of data for present study), concordance between patient and physician report of physician specialty was 93.7%.

The Committee on Human Research at The Johns Hopkins School of Hygiene and Public Health approved this study.

Participants were selected from the pool of enrollees in 16 MCOs using claims data or other central information sources. The inclusion criteria were age 18 years or older on September 1, 1993; enrollment in the MCO at the time of sampling; and 2 or more medical encounters (outpatient visits, ED visits, or hospitalizations) with a diagnosis of asthma (International Classification of Diseases, Ninth Revision, Clinical Modification³⁸ code 493.xx) between September 1991 and August 1993.

The sampling pool was divided into 2 strata: inpatient (at least 1 hospitalization or ED visit during the previous 24-month period) and outpatient (all asthma contacts in the outpatient setting). From each stratum, a random sample of 300 patients was selected from each MCO. If fewer than 300 patients had hospitalizations or ED visits, then the size of the outpatient stratum was increased to obtain a sample of at least 600 patients. This sampling strategy was selected to increase the number of participants with more severe disease. Individuals were excluded if they denied having asthma, had disenrolled, or were expected to disenroll before January 1, 1994.

In August 1993, 10 539 patients were sampled, of whom 8640 were eligible. Reasons for ineligibility included not having asthma (844 patients), disenrollment (839 patients), and other reasons (216 patients). Between September and December 1993, patients meeting eligibility criteria were mailed a questionnaire. To increase the response rate, patients who had not responded were sent a postcard reminder after 2 weeks, another questionnaire 2 weeks later, and followed up in another 2 weeks (if necessary) by a telephone call and an offer to complete the interview by telephone. A total of 6612 patients (77%) completed the survey. Data are not available to compare responses of patients completing the survey by mail vs telephone.

To help clinicians improve care, the National Heart, Blood, and Lung Institute sponsored the NAEPP^3,24 to develop guidelines for management of asthma based on evidence and a consensus of expert opinion. This analysis focused on 5062 patients (77% of patients who completed surveys) with incomplete control of asthma symptoms consistent with 1991 NAEPP guideline definitions of moderate or severe asthma. We used the 1991 NAEPP guidelines rather than those from 1997 for this study because the former existed at the time these data were collected.

The primary dependent variables of interest were indicators of consistency of care with various aspects of the NAEPP guidelines for patients with at least moderate asthma severity. For analytic purposes, we separated guideline recommendations into 5 domains: (1) medication, (2) self-management education, (3) control of factors related to asthma severity, (4) periodic assessment, and (5) asthma specialist care. In each of these domains, we selected 1 or more indicators of NAEPP-consistent care (Table 1). Table 2 presents the survey questions by care indicator.

The guidelines suggested a choice of 2 alternative medication regimens that we collapsed into a single indicator (NAEPP-recommended combination; Table 1). Because of the importance of airway inflammation in the pathogenesis of asthma and the effectiveness of ICS in reducing this inflammation,^39,40 patients were asked if they possessed an ICS. Among those who reported having an ICS, patients were also asked if ICS were used daily.

The NAEPP guidelines also recommended that an asthma specialist evaluate patients with moderate or severe asthma, so we included responses from patients regarding whether they had seen a specialist in the previous 1 year (yes; no, but would have preferred to; no, but did not need to).

Race and sex were the main independent variables. We restricted the study sample to whites and African Americans because there were too few patients of other races (<5% of all patients) to permit meaningful analyses. Other patient descriptors included age (18-35, 36-64, and ≥65 years), college education (yes [completed some college, a college graduate, or completed postgraduate work] vs no [high school graduate or less]), employment status (full-time or part-time work vs not working), smoking status (ever vs never), age of asthma onset (years), duration of asthma disease (current age minus age of onset, years), and history of atopy (allergies or hay fever).

The NAEPP-based classification of asthma severity and indications for asthma therapy are driven, in large part, by the level of asthma symptom control. Thus, information was collected about the frequency of several respiratory symptoms during the previous 4 weeks. We asked patients to report the frequency of cough, sputum production, chest tightness, wheezing, and shortness of breath using a 5-point scale (1, never; 2, once a week or less; 3, 2-3 times a week; 4, 4-5 times a week; and 5, daily). Patients were asked the frequency of nocturnal awakening due to asthma symptoms (1, never; 2, once; 3, 2-4 times; 4, 5-7 times, and 5, ≥8 times). Patients were also asked about the presence of asthma symptoms between attacks (1, no problem; 2, some symptoms on some days; 3, some symptoms on most days, requiring an inhaler for relief; and 4, symptoms most of the time). To account for symptom frequency in multivariate models while avoiding problems with collinearity, we combined the frequency of these respiratory symptoms into a global measure of asthma symptom severity (Asthma Symptom Index).³⁵ The index (range, 1-5) is the arithmetic mean of the patient-reported frequency for these respiratory symptoms. A higher score on the index indicates more frequent symptoms and has been shown to predict overuse of inhaled β₂-agonists.³⁶

Patients were asked to report the frequency of asthma attacks in the previous 4 weeks (1, not at all; 2, less than once a week; 3, 1 or 2 times a week; and 4, ≥3 times a week). For bivariate analyses, we tabulated the proportion of patients reporting 3 or more attacks a week (ie, threshold for defining moderate asthma severity based on frequency of attacks). We also collected information on utilization of acute health services during the previous year, including the number of ED visits for asthma and whether patients had been hospitalized for asthma.

Variables were examined using descriptive frequencies. Bivariate associations were measured using χ² tests for categorical variables and t tests or Wilcoxon rank sum tests for continuous variables. Simple and multivariate logistic regression models⁴¹ were constructed to determine whether race and sex were significantly associated with indicators of NAEPP-consistent care with and without adjusting for cross-sectional differences in age, college education, employment status, and Asthma Symptom Index. Age³⁶ and education⁴² were included because they have been reported to be independent predictors of inadequate pharmacotherapy. We included employment status to reduce potential confounding related to socioeconomic status and Asthma Symptom Index to account for reporting bias related to symptom frequency, respectively. The NAEPP recommendations for care are not based on factors such as age of asthma onset, duration of asthma, smoking history, or history of hospital admissions or ED visits, so we did not include them in multivariate analyses identifying independent predictors of care.

The general format for the multivariate logistic regression models for each care indicator is shown below: Log Odds (Care Indicator) = α + β₁(Sex) + β₂ (Race) + β₃(Race × Sex) + β₄(Age, 36-64 Years) + β₅(Age, 65 Years and Older) + β₆(College Education) + β₇(Employed Full-time or Part-time) + β₈ (Asthma Symptom Index).

We constructed separate models for the inpatient and outpatient sampling strata to determine if there were qualitative differences in the relationships between race, sex, and care. We included race-sex interaction terms (β₃[race × sex]) in the simple and multivariate logistic regression models to determine if the relationship between race (sex) and care was modified by sex (race). Finally, separate analyses were performed by MCO to determine if the relationships of race and sex to care were significantly different in the various MCOs. Model calibration was assessed using the Hosmer-Lemeshow goodness-of-fit test.⁴³ A 2-tailed P<.05 defined statistical significance for all analyses. Computations were performed with SAS version 6.07 software (SAS Institute, Cary, NC).

Of the 5062 patients with moderate or severe asthma symptoms, approximately 14% were African American and 72% were women (Table 3). Age was similar between the men and women. Whites were slightly older, on average, than African Americans (mean age, 44.6 vs 43.4 years; P<.001). A similar proportion of men and women were college educated (60.6% of all patients). Compared with African Americans, more whites reported college education. More African Americans than whites and more men than women were likely to be employed full-time or part-time. More whites than African Americans and more men than women reported having been a smoker. African Americans compared with whites (24.6 years vs 26.4 years; P = .02) and men compared with women (25.4 years vs 26.4 years; P = .03) reported slightly earlier age of asthma onset. Duration of asthma was not significantly different by race, but was slightly longer in men than women. More whites than African Americans and more women than men reported a history of allergies or hay fever.

There was a trend toward more frequent respiratory symptoms in whites compared with African Americans (Table 3). More whites reported frequent asthma attacks, while African Americans had substantially more ED visits and hospitalizations for asthma.

Overall, women reported slightly more frequent respiratory symptoms (Table 3). Although more men reported frequent asthma attacks, more women reported asthma-related ED visits and hospitalizations during the previous year.

A substantial proportion of all patients reported care that was inconsistent with NAEPP guidelines (Table 4 and Figure 1). The proportion of patients reporting indicators of care consistent with guidelines varied by MCO (data not shown), but where significant race or sex differences existed, the patterns of such differences were qualitatively similar (eg, favoring whites and/or men). For example, whites were more likely to report having an ICS in 11 of the 16 MCOs, with significant differences in 3 of 16 MCOs. However, African Americans were not significantly more likely to report having an ICS in any of the 16 MCOs. Thus, combined results from all MCOs are reported below.

In all 5 domains of asthma care, significantly fewer African Americans than whites reported care that was consistent with guideline recommendations. Greatest differences by race were seen in the daily use of an ICS (for patients with an ICS, 34.9% vs 54.4%; P = .001; African American vs white), education to avoid triggers (37.6% vs 53.6%; P = .001), and use of specialist care (28.3% vs 41.0%; P = .001). Race differences in specialist care did not appear to be driven by patient preferences for source of asthma care, since 30.3% of African Americans vs 17.2% of whites reported not seeing a specialist but wanting to (P = .001). There was a significant race-sex interaction for use of a NAEPP-recommended combination of asthma medications but not for other indicators of care. African American men were least likely to report a medication combination consistent with guidelines (65.8%; P<.01), whereas the rates were similar among other patients (African American women [78.9%], white women [78.2%], and white men [76.5%]).

Care indicators did not exclusively favor men or women. With the exception of medication and specialist domains, however, consistency of care with guidelines generally favored women. In the medication domain, men were less likely to possess an ICS but were more likely to use it daily if they had one (58.3% vs 49.6%; P = .001). Men were also more likely to have seen a specialist (43.1% vs 37.7%; P = .001; Figure 1) in the previous 12 months than were women. As in the analyses by race, lower use of specialist care in women compared with men did not seem to be related to patient preferences.

The relationships of race and sex to indicators of asthma care were similar in the inpatient and outpatient sampling strata. In addition, the race-sex interaction term was not significant in any of the multivariate models (P>.05 in all models; Table 5). Thus, we reported results of multivariate analyses in the combined inpatient-outpatient strata after excluding the race-sex interaction term. The Hosmer-Lemeshow goodness-of-fit test was not significant in any of these models, suggesting that the multivariate models were adequately calibrated.

After accounting for age, education, employment, and symptom frequency, there were no significant race (P>.99) or sex (P = .09) differences in the use of a medication regimen consistent with NAEPP recommendations for patients with moderate or more severe asthma. Also, men were less likely to report receiving instructions on the use of a peak flowmeter than women. Otherwise, results from bivariate and multivariate analyses were similar, with significantly fewer African Americans reporting care consistent with several components of the guidelines. As in the bivariate analyses, sex differences were mixed and generally favored women, with the exception of significantly greater daily ICS use and specialist care among men.

In this study of patients with moderate or severe asthma enrolled in managed care, African Americans were less likely than whites to report care that was consistent with a comprehensive array of guideline recommendations. By contrast, differences in asthma care by sex were small and, with the exception of daily ICS use and seeing an asthma specialist, tended to favor women. These differences in care by race and sex were largely unchanged after adjusting for age, education, employment status, and asthma symptom frequency. These findings suggest that even among patients with health insurance, differences in several aspects of medical management may contribute to race disparities, and, to a lesser extent, sex disparities in asthma outcomes.

Although patient-physician partnership in asthma care has been shown to improve outcomes and is emphasized in national guidelines,^3,24 African American patients were significantly less likely than whites to report education for self-management and avoidance of asthma triggers. Although we did not assess reasons for this race difference, there is increasing evidence in other settings to suggest that lower levels of partnership may, in part, represent cultural barriers to effective communication between patients and their physicians.^44,45 More studies are needed to better understand the basis for race differences in partnership for care among patients with asthma.

While there were no significant race or sex differences in whether patients reported use of 1 of 2 NAEPP-recommended medication regimens, African Americans and women were less likely to use an ICS daily. Similar race disparities in ICS use have been found in other studies evaluating care in MCOs.^8,23 To our knowledge, sex variations in the use of ICS have not been previously reported, perhaps reflecting selection bias (31% survey response rate⁸) and the limited ability to estimate patterns of ICS use with pharmacy data²³ in other studies.

Similar to our findings, previous studies in managed care populations found that whites were more likely than African Americans to receive care from asthma specialists.^23,46 In this study, we also found evidence of disparities in the use of asthma specialists by women. Lower use of specialist care in the previous 12 months was not readily explained by differences in patient preference, age, education, employment, frequency of respiratory symptoms, or health insurance, suggesting that there may be nonpatient barriers to specialist care among African Americans and women with asthma. Further studies are needed to determine whether physician referral patterns or health care system barriers to specialist care differ by race and sex.

Findings from this study have several implications. In this population with inadequate asthma symptom control, only 21.8% of African Americans, 39.3% of whites, 35.7% of women, and 39.6% of men reported daily use of ICS. Since ICS are the most effective long-term control medications for asthma,^39,40,47,48 efforts to increase regular ICS use should play a prominent role in strategies to improve asthma outcomes in all patients, but particularly in African Americans and women. In addition, the presence of lower levels of patient-physician partnership for African Americans with asthma suggests that efforts to reduce race inequities in asthma outcomes should extend beyond strategies based on improving medication use alone. Results from a few studies suggest that care provided by asthma specialists may be more likely than care from generalists to conform to national guidelines,⁴⁹ improve quality of life,⁴⁹ and reduce the number of ED visits for asthma exacerbations.^49,50 If the relationship between specialist care and improved outcomes is confirmed in other studies, our findings suggest that lower rates of specialist care among African Americans and women may contribute to race and sex disparities in asthma outcomes. To correct disparities in outcomes, it will be important to understand if differences in care assessed in this study were due to barriers attributable to providers (eg, offering care), patients (eg, acceptance of care), or health care systems (eg, availability of care).

Our study has several strengths. First, we evaluated care among adults enrolled in employer-based MCOs. Thus, the race- and sex-related inequities in asthma care we identified cannot be explained by lack of health insurance. Second, we evaluated the relationships of race and sex to care separately from those due to other demographic factors related to socioeconomic status (ie, age, education, and employment) and asthma symptom frequency. Third, we assessed care in a large patient population in several MCOs throughout the United States. Since MCOs provide care to a substantial proportion of patients in the United States and the relationships of race and sex to asthma care did not significantly vary by MCO, our findings may be generalizable to a large segment of patients with moderate or severe asthma receiving care in managed care settings.

This study also has potential limitations. Although there was a moderately high survey response rate (77%), results from this study may not be generalizable to all patients who were provided care in the various MCOs. These results may also not apply to patients with mild disease. We deliberately selected patients who had moderate or more severe asthma because we believe that there would be more agreement that the comprehensive strategy of care recommended by the guidelines would be indicated in patients with more severe disease. Moreover, these findings may not represent the experience of disadvantaged patient populations, such as those with lower prevalence of health insurance, college education, or employment. However, we believe that this nonrepresentative aspect of the study design is one of the greatest strengths of our findings, since we examined patterns of asthma care in a population where confounding due to socioeconomic factors was minimized. Another potential limitation is the risk of reporting bias related to the use of self-reported data. Although patients' recall of their medication use can be quite good,^51,52 some may have overreported adherence with medications,⁵³ resulting in biased estimates of asthma care. The purpose of this study, however, was to measure differences in asthma care by race and sex, rather than absolute levels of care. Nevertheless, if reporting bias varied by these demographic factors, our analyses may have minimized or exaggerated true differences in care. Results of studies evaluating differential reporting in other settings, however, have not been consistent (no race^54,55 or sex⁵⁶ bias, less bias⁵⁷ in African Americans compared with whites, and less bias⁵⁶ in whites compared with African Americans). Also, similar race differences in ICS use^8,23 and sex differences in peak flowmeter possession⁸ were reported in previous studies in managed care populations, including a study using pharmacy data.²³ For these reasons, we do not believe that race or sex differences in asthma care can be adequately explained on the basis of race or sex reporting bias alone. Other measures of medication use may be free from reporting bias, but they too have important limitations. For example, pharmacy data cannot differentiate between various patterns of use (eg, 2 puffs/day for 7 days/week vs 7 puffs/day for 2 days/week) and may underestimate care if patients received medications from other sources (eg, physician samples or from other pharmacies). While electronic monitors on multidose inhalers and peak flowmeters can accurately record patterns of use, they would not be feasible in large patient populations. Physicians' records can be used to determine if specific care was offered or discussed, but information on patterns of medication use may be incomplete and subject to patients' reporting biases. Importantly, none of these alternate study designs can provide information on other aspects of care we assessed, such as patients' perception regarding adequacy of self-management education. Despite adjustments for several factors related to socioeconomic status, there may have been residual confounding. Also, there may have been differences in other factors related to disease severity that were not adequately accounted for in the analyses. However, there is no universally accepted and validated measure of asthma severity, and NAEPP recommendations for asthma care do not specify that, in patients with moderate or more severe asthma symptoms, care should be different for patients with additional factors potentially related to severity.

In this well-educated population of patients with health insurance and moderate or severe asthma symptoms, we found race- and (to a lesser extent) sex-specific differences in consistency of care with national guidelines. This study raises serious concern for the quality of asthma care even among relatively advantaged patients with access to care. More research is necessary to understand the basis for these observed differences in care by race and sex. Our results suggest that a broad strategy that incorporates various components of the asthma guidelines may be necessary to improve outcomes in African Americans with asthma. To reduce sex disparities in outcomes, greater emphasis should be placed on daily ICS use for women with asthma.

Accepted for publication August 17, 2000.

This study was supported by a Clinical Research Trainee Award for Asthma by the CHEST Foundation of the American College of Chest Physicians and Glaxo Wellcome Inc (Dr Krishnan), an institutional training grant T32HL07534 from the National Heart, Lung, and Blood Institute (Dr Krishnan), and the Managed Health Care Association Outcomes Management System Project Consortium, Washington, DC. The MHCA study was coordinated by the Health Outcomes Institute, Minneapolis, Minn. Employer members of the consortium participating in the asthma project were Ameritech, Chicago, Ill; Becton Dickinson & Co, Franklin Lakes, NJ; Commonwealth of Virginia, Richmond; Digital Equipment Corp, Maynard, Mass; GTE Service Corp, Stamford, Conn; HealthTrust Inc, Nashville, Tenn; James River Corp, Richmond, Va; Marriott International Corp, Washington, DC; Procter & Gamble, Cincinnati, Ohio; Promus Companies, Memphis, Tenn; and Xerox Corp, Stamford, Conn. Managed care organizations participating in the consortium as partners of the above employers were Aetna Life Insurance Co, Chicago, Ill; Alliance Blue Cross Blue Shield, St Louis, Mo; Anthem Blue Cross and Blue Shield, Indianapolis, Ind; Blue Cross Blue Shield of Illinois, Chicago; Blue Cross Blue Shield of Massachusetts, Boston; Blue Cross Blue Shield of Rochester, Rochester, NY; Fallon Community Health Plan, Worcester, Mass; Harvard Pilgrim Health Care, Brookline, Mass; Intermountain Health Care, Salt Lake City, Utah; Kaiser Permanente/Ohio Region, Brooklyn Heights; Matthew Thornton Health Plan, Manchester, NH; The Prudential Health Care System, Atlanta, Ga; Trigon Blue Cross Blue Shield, Richmond, Va; United Health Care, Hartford, Conn; and USQA (US Healthcare), Bluebell, Pa.

We acknowledge Jonathan M. Samet, MD, MSc, and Noah Lechtzin, MD, MHS, for their thoughtful comments on previous drafts of the manuscript.

Corresponding author and reprints: Jerry A. Krishnan, MD, Division of Pulmonary and Critical Care Medicine, The Johns Hopkins University School of Medicine, 5501 Hopkins Bayview Cir, Room 4B.74, Baltimore, MD 21224 (e-mail: satish@welch.jhu.edu).