Chest ACCP Member Benefits
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Guest Access | Sign In via User Name/Password
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF) Free
Right arrow Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Article Archive
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via ISI Web of Science (4)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Bach, P. B.
Right arrow Articles by Bennett, C. L.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Bach, P. B.
Right arrow Articles by Bennett, C. L.
(Chest. 1999;115:1563-1569.)
© 1999 American College of Chest Physicians

The Relation Between Physician Experience and Patterns of Care for Patients With AIDS-Related Pneumocystis carinii Pneumonia*

Results From a Survey of 1,500 Physicians in the United States

Peter B. Bach , MD, MAPP; Elizabeth A. Calhoun , PhD and Charles L. Bennett , MD, PhD

From the Robert Wood Johnson Clinical Scholars Program (Dr. Bach), The University of Chicago, Chicago, IL; the Department of Medicine (Dr. Calhoun), Institute for Health Services Research and Policy Studies, Northwestern University, Chicago, IL; and the Department of Medicine (Dr. Bennett), Chicago VA Healthcare System, Chicago, IL.

Correspondence to: Peter B. Bach, MD, MAPP, Department of Epidemiology and Biostatistics, Box 44, Memorial Sloan-Kettering Cancer Center, 1275 York Ave, New York, NY 10021; e-mail: bachp@biosta.mskcc.org


   Abstract
TOP
 Abstract
Introduction
 Materials and Methods
 Results
 Discussion
 Appendix A
 Appendix B
 References
 
Study objectives: To determine whether physician experience and specialty influence the approach to care of AIDS patients with pneumonia, we surveyed physicians about their management of possible Pneumocystis carinii pneumonia (PCP) infection.

Design, setting, participants: A postal survey was sent to a random sample of 1,500 internists and family physicians in the United States drawn from the American Medical Association master file who were identified by a pharmaceutical marketing company as having written prescriptions for AIDS-related agents in the previous year.

Measurements and results: The survey had a 53% response rate. Physicians more experienced in AIDS care were more likely to advocate diagnostic bronchoscopy over initiation of empiric anti-PCP therapy for HIV-infected patients with undiagnosed pulmonary infiltrates (odds ratio [OR], 1.4 for a patient with mild severity of illness [p = 0.02]; OR, 1.7 for a severely ill patient [p < 0.001]). Physician specialty and fee-for-service reimbursement were independently associated with higher rates of bronchoscopy, with internists favoring bronchoscopy more frequently than family physicians. High-experience providers and internists also predicted better clinical outcomes for the hypothetical patients.

Conclusions: Our findings extend the observations about HIV experience and PCP prophylaxis to the setting of diagnosis and treatment. Physicians with higher levels of experience with AIDS, internists, and physicians reimbursed as fee-for-service providers are more likely to support diagnostic confirmation of PCP than empiric treatment approaches.

Key Words: AIDS • outcome and process assessment • physician practice patterns • pneumonia • postal survey • prognosis


   Introduction
TOP
 Abstract
 Introduction
Materials and Methods
 Results
 Discussion
 Appendix A
 Appendix B
 References
 
The most common reason for hospitalization and one of the most common causes of death for persons with AIDS has been Pneumocystis carinii pneumonia (PCP).1 Although patients with PCP have higher survival rates when they receive care at hospitals with high levels of AIDS experience, some researchers have hypothesized that variation in mortality rates is caused by selective referral of healthier patients to high-experience hospitals.2 Others have suggested that high-experience hospitals are likely to provide better care, presumably through more efficient use of medical resources, better organization of medical services, or use of more medical re-sources.3 4

Patterns of appropriate care for patients with suspected cases of PCP have been identified previously. A consensus panel indicated that early use of diagnostic tests such as bronchoscopy was one of the most important elements of good process of care, primarily because of concern that other pathogens such as pulmonary tuberculosis and community-acquired bacterial infection may go unrecognized.5 Among HIV-infected individuals with suspected cases of PCP who were treated in Chicago, IL, Los Angeles, CA, and Miami, FL, between 1987 and 1990, persons who underwent bronchoscopy within 2 days of admission were one third less likely to die in-hospital than others, even after adjustment for differences in the severity of illness. Although studies have found that physicians with higher levels of experience with AIDS are more likely to use appropriate antiretroviral therapies and anti-PCP prophylaxis and that their patients have longer survival rates, to our knowledge, no prior study has included information on variations in physician practice patterns for severe opportunistic infections such as HIV-related PCP.6 7

In this study, we sought to better understand the preferences of individual physicians to pursue diagnostic confirmation vs empiric treatment for patients with suspected HIV-related PCP. With the assistance of a pharmaceutical marketing company database, we targeted a broad range of physicians who varied with respect to specialty training and experience with HIV-infected individuals. The survey addressed two general types of questions about variations in care of suspected PCP. Are physicians with higher levels of HIV experience more likely to pursue diagnostic confirmation than their less-experienced colleagues? Do high-experience physicians predict better outcomes for persons with HIV-related PCP? We also wondered whether physicians who reported being reimbursed as fee-for-service would favor bronchoscopy more than their peers, as suggested by previous research into HIV care.8


   Materials and Methods
TOP
 Abstract
 Introduction
 Materials and Methods
Results
 Discussion
 Appendix A
 Appendix B
 References
 
Subjects and Survey Instrument
The sample of 1,500 subjects was drawn from a pool of 11,000 board-certified family practitioners and internists in the United States who both had written prescriptions for medications used in the treatment of HIV disease during the time period from February 1996 through February 1997 and had addresses in the American Medical Association (AMA) master file. The pool was generated from a physician database maintained by Pharmaceutical Marketing Services Inc (Phoenix, AZ) that captures physician prescribing patterns. The database does not count refills, but also does not distinguish between the medications prescribed, whether the prescriptions go to the same patient or different patients, or whether the prescriptions are renewals of a medication that the patient is already taking.

Physicians were divided into equal-sized quintiles of prescribing patterns, with the first quintile containing physicians who had written between 1 and 2 prescriptions for AIDS-related drugs in the previous year (for a list of the agents, ); the second, 3 to 6 prescriptions; the third, 7 to 16; the fourth, 17 to 52; and the fifth, 53 to 2,314. We randomly selected 500 physicians to receive the survey from each of the top three quintiles to maximize the likelihood that our subjects were active in the care of patients with HIV infection. We have described the utility of this sampling strategy elsewhere.9

Subjects were mailed a five-page survey booklet requiring approximately 7 min to complete, given a $1 incentive, and assured that participation was voluntary and that responses were confidential. Those who did not respond within 50 days were sent a second copy of the survey. This research was approved by the Institutional Review Board at the University of Chicago. Every subject received the same survey, with the exception that the patient's risk factors were randomly assigned, so that a particular recipient might be informed that one of the patients had acquired HIV through a transfusion, whereas another recipient might be informed that the same patient had acquired HIV through injection drug use.

The survey elicited the following: (1) responses to questions about the use of diagnostic bronchoscopy in the care of two hypothetical patients with possible PCP; (2) responses to questions about clinical experience with the care of AIDS patients; (3) professional and demographic information, including a response to the question "How would you define most of your clinical reimbursement" followed by choices of "fixed salary," "fee-for-service," "capitation," and "other." Before distribution, the survey instrument was reviewed by 2 experts in survey administration and composition, and 6 experts in infectious diseases and pulmonary medicine, and was pretested on 30 internists and 5 family physicians.

Measures of Practice
Two scenarios described patients with HIV infection who had undiagnosed pulmonary processes and were at risk for PCP. A brief history, list of symptoms, chest radiographic findings, and arterial blood gas analysis were included in each description. After reading each scenario, subjects were told that noninvasive tests for PCP infection were unavailable and were asked a number of questions about the patient's clinical situation, including whether they thought bronchoscopy was indicated. The first scenario described an outpatient with an indolent process, cough, and infiltrates, and subjects were asked to choose either diagnostic bronchoscopy or empiric treatment with an appropriate antibiotic (trimethoprim-sulfamethoxazole double-strength, two pills, three times each day).

The second scenario was a multipart case history. Section 1 (shown in ) of the scenario described a patient 3 days after admission to the ICU for respiratory failure. After this section, subjects were asked, on a five-point Likert scale, their preference for bronchoscopy. Section 2 of this latter scenario focused on day 7 of ICU care, with the patient having renal insufficiency and circulatory failure.

Measures of Prognostication
The subjects were asked to estimate the likelihood that the mildly ill patient (scenario 1) would recover to the level of health he had before his illness and to estimate, after the first and second parts of the second scenario, the likelihood that the severely ill patient would survive to hospital discharge.

Professional, Demographic, and Attitudinal Variables
We combined data on the subjects from three sources: responses to survey questions about the subject's practice setting, reimbursement structure, and exposure to AIDS patients and critically ill patients in the past year; demographic information about the subject's specialty training and demographics from the AMA master-file; and recent prescribing history for AIDS-related therapeutics from Pharmaceutical Marketing Services Inc.

Statistical Analysis
All p values are two-sided. The five-point Likert scale reflecting preference for bronchoscopy in scenario 2 was dichotomized so that any response more affirmative than the neutral middle position was judged as a preference for bronchoscopy. Physicians were categorized as internists or family physicians based on their specialty designation (field "BOARD CODE 1") in the AMA master file. Physicians were characterized as less experienced if they had written <= 16 prescriptions (lowest third of all physicians surveyed) for AIDS-related therapeutics in the previous year. When asked to estimate likelihoods of outcomes, some subjects answered with a range (eg, 70 to 85%). The midpoint of the range was then used for analysis.

Pearson {chi}2 statistics were generated to judge the effect of the subject's specialty, experience, type of reimbursement, and patient's risk factor on the choice to perform bronchoscopy, and to compare responders with nonresponders on dichotomized demographic variables such as gender, specialty (internal medicine or family medicine), and country of origin (United States or non-United States). Estimates of survival probability and ages of respondents and nonrespondents were compared using the Student's t test. Predictors of preference for bronchoscopy were combined using multiple logistic regression. Predictors of likelihoods of positive outcomes were combined using multiple linear regression. Statistical analyses were performed using appropriate software (STATA; Stata Corporation; College Station, TX).


   Results
TOP
 Abstract
 Introduction
 Materials and Methods
 Results
Discussion
 Appendix A
 Appendix B
 References
 
Sample
Of the 1,500 physicians surveyed, we received 900 responses. Of these, 115 (8%) were either returned cover letters, undeliverable, incomplete, blank, or from subjects who reported that they had not seen a patient with AIDS in the previous year. We were left with 785 responses in which the subject was eligible and answered the questions after each scenario, yielding a 53% usable response rate. Respondents did not differ from nonrespondents with respect to age or specialty. Respondents were more likely to be female than nonrespondents, (odds ratio [OR], 1.3; p = 0.04), and to have been born in the United States (OR, 1.5; p = 0.002). The mean age of the respondents was 47 years old; 80% of the respondents were male; 74% of the respondents were trained in internal medicine, and the remainder were trained in family medicine. Of those trained in internal medicine, 304 were general internists (52%), with the bulk of the remainder specializing in infectious diseases (20%), pulmonary diseases (5%), and oncology (5%). Thirty-nine percent practiced in community hospitals, and 18% practiced in university or teaching hospitals. The respondents spent an average of 88% of their time engaged in clinical duties. Stated experience with the care of patients with AIDS or HIV infection and experience caring for patients with PCP correlated closely with the number of prescriptions written by the physician in the previous year (Kruskal's {gamma}, 0.66).10

Preference for Bronchoscopy as a Diagnostic Aid
In the first scenario, the subjects could choose either to treat an outpatient with empiric antibiotics for PCP or to proceed to diagnostic fiberoptic bronchoscopy. In the second scenario, the subjects were asked their strength of preference for bronchoscopy in a critically ill AIDS patient with undiagnosed infiltrates. Overall, 32% of the respondents preferred bronchoscopy over empiric antibiotics in the first scenario; in the second scenario, 64% of respondents favored bronchoscopy over empiric treatment.

Table 1 reflects three findings. The first is that physicians with more HIV experience were much more likely to prefer bronchoscopy in both scenarios than were those who cared for relatively few HIV-infected patients: 34% vs 26%, respectively, for the first scenario (p = 0.02); and 68% vs 55%, respectively, for the second scenario (p = 0.004). Table 1 also indicates that internists, compared with family physicians, were much more likely to favor bronchoscopy in both scenarios: 35% vs 23%, respectively, for the first scenario (p = 0.004); and 66% vs 56%, respectively, for the second scenario (p = 0.01). Physicians who reported being primarily reimbursed through fee-for-service mechanisms also were much more likely to favor bronchoscopy in both scenarios: 37% vs 27%, respectively, for the first scenario (p = 0.004); and 76% vs 53%, respectively (p < 0.001).


View this table:
[in this window]
[in a new window]

  		Table 1. Preference for Bronchoscopy in Scenarios 1 and 2: Comparison of the Effects of Experience, Specialty, and Reimbursement*

 
We also examined these predictors with multiple logistic regression, controlling for other possible predictors of bronchoscopy: practice setting, university affiliation, and urban vs nonurban location. None of these had a significant impact on the model (data not shown). The results of the analyses that included specialty, experience, and reimbursement are displayed in Table 2 . Of note, the differences demonstrated in Table 1 between experienced and inexperienced physicians (p < 0.01 for both scenarios), between internists and family physicians (p < 0.09 for both scenarios), and between fee-for-service and other physicians (p < 0.01) remained strong in both scenarios. There were no significant differences (data not shown) in preferences for bronchoscopy between physicians who were subspecialists in infectious diseases compared with all other internists (p = 0.29 for first scenario; p = 0.73 for second scenario).


View this table:
[in this window]
[in a new window]

  		Table 2. Multiple Predictors of Preference for Bronchoscopy Based on Experience, Specialty Training, and Type of Reimbursement*

 
The patient's risk factor had no discernible impact on the preference for bronchoscopy in the scenario with the mildly ill patient, in which subjects favored bronchoscopy for 29% of transfusion-acquired AIDS patients, 34% of injection drug use-acquired AIDS patients, and 32% for homosexual sex-acquired AIDS patients (p = 0.57). In contrast, there were significant differences in the preference for bronchoscopy in the scenario with the severely ill patient, with physicians favoring bronchoscopy in 57% of transfusion-acquired AIDS patients, 71% of injection drug use-acquired AIDS patients, and 63% for homosexual sex-acquired AIDS patients (p = 0.004).

Estimate of the Probability of Favorable Outcomes
As part of each scenario, subjects were asked to estimate the probability of particular events occurring. After the scenario with the mildly ill patient, subjects were asked to estimate the probability that the described patient would return to his previous state of health after his acute illness. As part of the scenario with the severely ill patient, subjects were asked to estimate the likelihood of survival to hospital discharge both at 3 days and at 7 days after ICU admission. Overall, physicians estimated a (mean ± SD) 79 ± 20% likelihood that the mildly ill patient would recover to his previous state of health. Physicians estimated that after 3 days in the ICU, the severely ill patient had a 45 ± 22% likelihood of survival. After 7 days, the estimated probability of survival had declined to 16 ± 14%.

Table 3 displays the differences in these estimates by HIV experience and specialty. When compared with experienced physicians, inexperienced physicians estimated a lower likelihood of full recovery for the mildly ill patient (72% vs 81%, respectively [p < 0.001]), and lower likelihoods of survival for the severely ill patient after 3 and 7 days of ICU care (40% vs 48%, respectively, after 3 ICU days [p < 0.001]; 14% vs 18%, respectively, after 7 ICU days [p = 0.01]). When compared with internists, family physicians consistently estimated lower likelihoods of good outcomes for both patients: for the first patient's chance of full recovery, 74% vs 80%, respectively (p < 0.01); for the second patient's chance of survival at each point, 40% vs 47%, respectively (p < 0.002); and 14% vs 17%, respectively (p = 0.09). In contrast, differences in reimbursement were not associated with differences in prognostication. We examined, with multiple linear regression on outcome prediction, the effects of specialty and experience. With the exception of the second prediction in scenario 2, in which the impact of being an internist on prediction was no longer statistically significant (p = 0.14), the impact of specialty and experience remained significant in all analyses (p < 0.02 for all predictors, data not shown).


View this table:
[in this window]
[in a new window]

  		Table 3. Expectations of Positive Outcomes in Scenarios 1 and 2, Stratified by Physician Experience With HIV Care, Physician Specialty, and Reimbursement*

 

   Discussion
TOP
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
Appendix A
 Appendix B
 References
 
In this study, we identified a large number of internists and family physicians directly involved in the care of HIV-infected individuals in the United States through use of a pharmacy database and surveyed them to assess whether there were differences in care or differences in expectations of care associated with several physician or patient factors. We found that physicians with higher levels of HIV experience were more likely to pursue diagnostic confirmation of PCP and had expectations of better outcomes, whereas physicians with lower experience levels were more likely to initiate empiric treatment and had expectations of poorer outcomes. We also found that internists, more than family physicians, and physicians reimbursed under fee-for-service incentives were more likely to advocate bronchoscopy over treatment with empiric antibiotics. To our knowledge, our study is the first to report these different approaches to care for patients with HIV-related PCP, one of the most common and costly opportunistic infections seen in persons with AIDS.

Improvements in outcomes for PCP patients have been reported during the past decade, with short-term survival rates for severely ill patients improving from 55% in the period from 1987 to 1990 to > 80% in 1996, and the short-term survival rates for mildly ill patients improving from 85% to > 95%.1 Although improved outcomes have been attributed to changes in both diagnostic strategies as well as to the use of adjunctive corticosteroids, the routine use of bronchoscopy in the diagnosis of pulmonary infiltrates in an HIV-infected patient remains controversial. There is ample evidence that bronchoscopy is an effective diagnostic tool and that AIDS patients with pulmonary infiltrates who receive bronchoscopy are more likely to receive an accurate diagnosis for their pulmonary process. On the other hand, the risks of initial empiric therapy, especially among less severely ill individuals, appear to be low, raising questions about both the optimal initial approach to care and the cost-effectiveness of bronchoscopy.11 12 13 14 15 16 17 We are not aware of any authors who recommend empiric therapy for critically ill patients. Consistent with this reasoning, preferences for the use of diagnostic bronchoscopy were higher for the scenario with the severely ill patient. About one third of physicians with higher HIV experience and 26% of those with lower HIV experience supported the use of a diagnostic bronchoscopy for the mildly ill patient, whereas more than two thirds of the higher HIV-experience physicians and 55% of lower HIV-experience physicians supported this approach for a severely ill patient.

Early in the AIDS epidemic, it was believed that HIV care could be effectively delivered by any primary care provider.18 Although the early studies showed that hospitalized patients received better HIV care and had better survival rates when treated at hospitals with higher levels of experience, the patterns and intensity of resource use did not differ from those at lower-experience facilities.3 19 It was believed by some that high-experience hospitals provided care for less severely ill patients and may not have provided better HIV care.2 18 As the complexity of AIDS increased, a study published in 1996 demonstrated that survival with HIV disease was significantly better in patients whose physicians had experience in HIV care and were able to provide appropriate outpatient antiretroviral therapies and prophylaxis for PCP and other severe opportunistic infections.6 This finding, coupled with the rapidly growing number of new therapeutic drugs, has changed the paradigm such that it is now recommended that HIV or AIDS primary care be received from experienced providers.20 21

Because AIDS is a relatively new disease for which no formal subspecialty board examinations have been instituted, physicians who provide care for HIV-infected individuals vary in their prior training experiences. After reports that HIV infection was a generalist illness, physicians with broad training backgrounds developed active practices for the care of HIV-infected individuals.18 In this study, we found that general internists and infectious disease specialists appeared to systematically differ from family physicians in their approach to HIV-related PCP care. Consistent with the finding by Curtis et al22 with a simulated patient, the internists and infectious disease specialists were more likely than family physicians to pursue diagnostic confirmation of PCP as an alternative to initiating empiric treatment.

Interestingly, the internists and experienced providers also predicted likelihoods of favorable outcomes higher than those of family physicians and physicians with low HIV experience. Paauw et al23 and Turner et al7 have found similar results in evaluating variations in care for HIV-infected individuals—generalists are likely to provide a lower quality of care. Not surprising is our finding that physicians reimbursed under fee-for-service systems were more likely to advocate bronchoscopy but demonstrated no differences from their peers about the expectation of outcome, adding evidence to the argument advanced by others that this incentive operates independently of other factors.8

This study has several limitations. First, our results reflect physicians' attitudes about hypothetical clinical situations and may not accurately reflect true clinical practice. We attempted to control for the large variations in clinical presentations of individuals with HIV-related PCP by varying the risk-group description to include IV drug users, homosexual and bisexual men, and persons with transfusion-acquired AIDS and by including scenarios of both mildly ill and severely ill patients. However, we found that none of these variations consistently affected our results. Second, although we found no meaningful evidence that our 785 responders do not represent our entire sample of 1,500, and a response rate of 53% is consistent with other published surveys of physicians,24 it is possible that some results do not represent the attitudes of the entire pool of physicians. Third, we queried physicians whom we identified through their prescription-writing patterns and may have missed physicians, such as intensivists, hospitalists, and physicians who attend on dedicated AIDS units, whose practices meaningfully influence the care of patients with AIDS in the United States.

Despite these limitations, our results support the growing body of literature that identifies differences in AIDS care and outcome according to the HIV experience level of individual physicians. Recently, the complexity of HIV care has increased dramatically with the development of highly active antiretroviral therapies, widespread use of prophylactic treatments for PCP and atypical mycobacterial infections, and many new treatments for serious opportunistic infections. In conjunction with the findings of others, our study suggests a mechanism by which outcomes for HIV-infected individuals improve when they are cared for by physicians with a high degree of experience with AIDS—more experienced providers are more optimistic about outcomes and are more aggressive in pursuing diagnostic confirmation of a possible opportunistic infection.


   Appendix A
TOP
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
 Appendix A
Appendix B
 References
 
AIDS Treatment-Related Agents
Physicians were identified as caring for patients with HIV infection if they had written prescriptions for stavudine, zalcitabine, saquinavir, rifabutin, nevirapine, ritonavir, zidovudine, atovaquone, didanosine, lamivudine, or pentamidine.


   Appendix B
TOP
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
 Appendix A
 Appendix B
References
 
Sample Scenario
John is a 26-year-old man with AIDS whose HIV risk factor is homosexual sex. He presents to the emergency department after having a seizure. An MRI reveals two round, ring-enhancing lesions in his left frontal lobe. He is admitted to the hospital and begins receiving phenytoin and empiric treatment for Toxoplasma gondii infection. After a few days, he experiences dyspnea and hypoxemia, and requires intubation. He has never had PCP and is not receiving PCP prophylaxis. The results of various tests include the following:

Arterial blood gases: fraction of inspired oxygen, 50%; pH, 7.41; PCO2, 35 mm Hg; PaO2, 74 mm Hg.

Chest radiograph: left apical cavitary lesion, diffuse bilateral fluffy infiltrates.

CD4 count: 25 cells/µL.

Respiratory rate: 35 breaths/min.

Suction catheter sputum: mixed Gram-negative organisms; no P carinii.

John should have a bronchoscopy in the next 24 h (please circle one):

Strongly disagree

Somewhat disagree

Neither agree nor disagree

Somewhat agree

Strongly agree


   Footnotes
 
Supported by The Robert Wood Johnson Foundation and a grant from the National Institutes of Health (RO1 DA10628–02). Pharmaceutical Marketing Services Incorporated (Phoenix, AZ) provided the physician mailing list and the prescription records free of charge for unrestricted use.

Abbreviations: AMA = American Medical Association; OR = odds ratio; PCP = Pneumocystis carinii pneumonia

Received for publication October 8, 1998. Accepted for publication January 12, 1999.


   References
TOP
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
 Appendix A
 Appendix B
 References
 

  1. Bennett, CL, Curtis, JR, Achenbach, C, et al (1996) US hospital care for HIV-infected persons and the role of public, private, and Veterans Administration hospitals. J Acquir Immune Defic Syndr Hum Retrovirol 13,416-421[ISI][Medline]
  2. Shapiro, MF, Greenfield, S (1992) Experience and outcomes in AIDS [editorial]. JAMA 268,2698-2699[CrossRef][ISI][Medline]
  3. Bennett, CL, Garfinkle, JB, Greenfield, S, et al (1989) The relation between hospital experience and in-hospital mortality for patients with AIDS-related PCP. JAMA 261,2975-2979[Abstract]
  4. Stone, VE, Seage, GR, III, Hertz, T, et al (1992) The relation between hospital experience and mortality for patients with AIDS. JAMA 268,2655-2661[Abstract]
  5. Bennett CL, Mathews C, Kosecoff J. Results of a consensus panel on process of care for patients with Pneumocystis carinii pneumonia [abstract]. The Fifth International Conference on AIDS; Montreal, Canada; 1989; 284A
  6. Kitahata, MM, Koepsell, TD, Deyo, RA, et al (1996) Physicians' experience with the acquired immunodeficiency syndrome as a factor in patient's survival. N Engl J Med 334,701-706[Abstract/Free Full Text]
  7. Turner, BJ, McKee, L, Fanning, T, et al (1994) AIDS specialist versus generalist ambulatory care for advanced HIV infection and impact on hospital use. Med Care 32,902-916[ISI][Medline]
  8. Horner, RD, Bennett, CL, Rodriguez, D, et al (1995) Relationship between procedures and health insurance for critically ill patients with Pneumocystis carinii pneumonia. Am J Respir Crit Care Med 152,1433-1444[ISI][Medline]
  9. Bach, PB, Calhoun, EA, Bennett, CL (1999) Identifying providers of HIV-care for a mail survey using a prescription tracking data base. J Clin Epidemiol 52,147-150[CrossRef][ISI][Medline]
  10. Agresti, A (1990) Categorical data analysis. John Wiley New York, NY.
  11. Huang, L, Hecht, FM, Stansell, JD, et al (1995) Suspected Pneumocystis carinii pneumonia with a negative induced sputum examination: is early bronchoscopy useful? Am J Respir Crit Care Med 151,1866-1871[Abstract]
  12. Tu, JV, Biem, HJ, Detsky, AS (1993) Bronchoscopy versus empirical therapy in HIV-infected patients with presumptive Pneumocystis carinii pneumonia: a decision analysis. Am Rev Respir Dis 148,370-377[ISI][Medline]
  13. Ognibene, FP, Shelhamer, J, Gill, V, et al (1984) The diagnosis of Pneumocystis carinii pneumonia in patients with the acquired immunodeficiency syndrome using subsegmental bronchoalveolar lavage. Am Rev Respir Dis 129,929-932[ISI][Medline]
  14. Masur, H (1993) Prevention and treatment of Pneumocystis pneumonia [review]. N Engl J Med 327,1853-1860[ISI][Medline]
  15. Hardy, WD (1995) Treatment and prevention of PCP and multiple opportunistic pathogen prophylaxis strategies. Improv Manag HIV Dis 3,20-24
  16. Golden, JA, Hollander, H, Stulbarg, MS, et al (1986) Bronchoalveoloar lavage as the exclusive diagnostic modality for Pneumocystis carinii pneumonia: a prospective study among patients with acquired immunodeficiency syndrome. Chest 90,18-22[Abstract/Free Full Text]
  17. Masur, H, Shelhamer, J (1996) Empiric outpatient management of HIV-related pneumonia: economical or unwise? Ann Intern Med 124,451-453[Free Full Text]
  18. Northfelt, DW, Hayward, RA, Shapiro, MF (1988) The acquired immunodeficiency syndrome is a primary care disease [review]. Ann Intern Med 109,773-775
  19. Cotton, DJ (1989) Improving survival in acquired immunodeficiency syndrome: is experience everything? [editorial] JAMA 261,3016-3017[CrossRef][ISI][Medline]
  20. Zuger, A, Sharp, VL (1997) `HIV specialists': the time has come. JAMA 278,1131-1132[CrossRef][ISI][Medline]
  21. Holmes, WC (1997) Quality of care in HIV/AIDS care: specialty or experience related? J Gen Intern Med 12,195-197[CrossRef][ISI][Medline]
  22. Curtis, JR, Paauw, DS, Wenrich, MD, et al (1995) Ability of primary care physicians to diagnose and manage Pneumocystis carinii pneumonia. J Gen Intern Med 10,395-399[ISI][Medline]
  23. Paauw, DS, Wenrich, MD, Curtis, JR, et al (1995) Ability of primary care physicians to recognize physical findings associated with HIV infection. JAMA 274,1380-1382[Abstract]
  24. Asch, DA, Jedrziewski, MK, Christakis, NA (1997) Response to rates to mail surveys published in medical journals. J Clin Epidemiol 50,1129-1136[CrossRef][ISI][Medline]



This article has been cited by other articles:


Home page
 Am. J. Respir. Crit. Care Med.Home page
J. M. BECK, M. J. ROSEN, and H. H. PEAVY
Pulmonary Complications of HIV Infection . Report of the Fourth NHLBI Workshop
Am. J. Respir. Crit. Care Med., December 1, 2001; 164(11): 2120 - 2126.
[Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF) Free
Right arrow Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Article Archive
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via ISI Web of Science (4)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Bach, P. B.
Right arrow Articles by Bennett, C. L.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Bach, P. B.
Right arrow Articles by Bennett, C. L.

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS