The effect of parental expectations on treatment of children with a cough: a report from ASPN.
In more routine clinical situations, physicians probe for information, but generally do not consult the patient before making a diagnosis or prescribing therapy. Both patients and physicians consider those to be the physicians responsibilities. Physicians generally have not been taught to consider the patient's perception of the problem or his or her expectations for treatment in these ordinary encounters.
Several authors, however, have suggested that communication between physician and patient be improved,[2-5] With emphasis on the physician listening to what the patient thinks is wrong and what should be done about it. As Press noted in 1984, patients approach their physicians with fully developed "explanatory models" of their illness, complete with diagnosis and treatment. Press suggested that physicians "develop a sensitivity" to patients' explanatory models. The extent to which physicians should allow those models to influence their diagnostic and therapeutic decisions, however, is uncertain.
In a previous retrospective study, examined how physicians diagnose bronchitis in a child with a cough. A history of sputum production, the finding of rales or rhonchi on examination, and a past history of lower respiratory tract illness were associated with a diagnosis of bronchitis. However, none of these symptoms were found in 24 of the 54 children diagnosed with bronchitis in that study. Because a diagnosis of bronchitis was closely associated with the use of an antibiotic, even among the 24 children without sputum production or rales, we speculated that the diagnostic label may, in some cases, follow the decision to treat instead of leading to it. If this is true, then factors influencing the decision to treat, such as parental expectations for an antibiotic, may affect the diagnosis of bronchitis.
Whether antibiotics should be used to treat acute bronchitis in children is open to debate. Most reviewers, [8,9] citing careful studies of etiologic agents that indicate that viruses are the predominant causes of acute bronchitis,[10,11] assert that antibiotic treatment is of little or no value, but they cite no controlled clinical trials of antibiotic treatment for bronchitis in children. Others suggest that there are times when antibiotic therapy might be appropriate; furthermore, antibiotics are commonly prescribed, not only in the United States, but also in Norway and the Netherlands.
The present study does not address the question of whether antibiotics should be used to treat children with acute bronchitis. Rather, it assesses the extent to which physicians incorporate parents' expectations into their medical decision making in ordinary outpatient practice. Specifically, if a physician perceives a parental expectation that an antibiotic will be prescribed, how does that affect the probability that a diagnosis of acute bronchitis will be made in a child with a cough?
The Ambulatory Sentinel Practice Network (ASPN) is an office-based research network of primary care practices in the United States and Canada. Forty-four practices participated in this study. Data were collected prospectively from October 22, 1990, to January 20, 1991. All children, newborn to 14 years of age, visiting an ASPN practice with a chief complaint of cough of up to 1 month's duration were eligible.
Data were reported by clinicians on standard data cards after the office visit. Data collected included the patient's sex, age, and duration of the illness (in days). Dichotomous variables summarized the patient's history and physical findings, including a history of fever, sputum production, presence of a smoker in the household, enrollment in day care, a history of allergies, a finding of rales on examination, a finding of wheezes, and whether a chest radiograph was obtained. Physicians recorded diagnosis and treatment by checking one or more boxes on the form, and indicated follow-up plans using an ordinal scale ranging from O ("well; no follow-up needed") to 4 ("hospitalized").
The clinicians were also instructed, "Indicate [on the data card] whether you sense an expectation by the patient's parent or guardian to prescribe an antibiotic." The study protocol did not specify how the physician was to "sense" that expectation. The physician's perception may have come from an explicit statement by the parent or from some nonverbal cue. An assessment of the origins of the clinicians' perceptions was not done because of the concern that such an assessment might alter the perceptions themselves or the reporting of them. The study did not require the physicians to make any changes in their usual clinical protocol. Clinicians made their diagnoses and treatment decisions in their usual manner.
Since the variables used in this study were dichotomous, ordinal, or skewed, chi-square and Mann-Whitney U tests were used for univariate comparisons. Where a measure of effect size was desired, the relative risk (RR) and its 95% confidence interval (CI) were calculated. Logistic regression, used for multivariate analysis, provides odds ratios that, because of the high prevalence of bronchitis within this study, do not accurately estimate RRs. Adjusted odds ratios, however, differed very little from the unadjusted values; therefore, only the unadjusted RRs are reported.
A total of 1398 patients were entered in the study by 44 practices; 47% were girls and 53% were boys. The children ranged in age from infancy to almost 15 years, with an average of 4.8 years (standard deviation [SD] 3.9 years). The age distribution was skewed toward younger children, the mode being 1-year-olds.
Most patients were not seriously ill. Only 12 (0.9%) were hospitalized and 46 (3.3%) were asked to return to the office within 2 days; 869 (62.2%) were believed to require no scheduled follow-up visit. Three fourths of the children had been coughing for a week or less; the mean duration was 6.8 days (SD 6.2 days), with a skewed distribution and a mode of 2 days.
A diagnosis of bronchitis was made in 465 children 33%), a viral upper respiratory tract infection (URTI) in 494 (35%), a bacterial infection (such as otitis media) in 379 (27%), asthma in 129 (9%), other allergic diseases in 52 (4%), pneumonia in 78 (6%), and croup in 30 (2%). More than one diagnosis was made for 268 children (19%). The most common combinations were bronchitis with bacterial URTIs (70 children), viral and bacterial URTIs (41), and asthma with bronchitis (38). Of the 20 patients with three diagnoses, 19 had a diagnosis of bronchitis.
The physician sensed that the parent expected an antibiotic in 215 cases (15.4%); 30 of the 44 practices reported at least one such instance. Of the 14 practices enrolling 30 or more children, 12 reported at least one instance of the physician perceiving that the parent expected an antibiotic prescription.
If the physician perceived that the parent expected a prescription for an antibiotic, the likelihood that a diagnosis of bronchitis would be made doubled. Practitioners sensed this parental expectation during office visits with 126 of 465 children in whom bronchitis was diagnosed (27.1%), but in only 89 of 933 children in whom it was not (9.5%). Only the finding of rales on physical examination was more strongly correlated with the diagnosis of bronchitis (Table 1). A history of sputum production, wheezes on examination, a history of fever, and enrollment in day care were also statistically associated with a diagnosis of bronchitis; for each of these findings, the relative risk of making a diagnosis of bronchitis, when that factor is present, is given in Table 1. Children diagnosed with bronchitis had been coughing longer, but the difference was small (7.3 vs 6.5 days, P = .03). The effect of parental expectations on the diagnosis of bronchitis was essentially unchanged when the influence of other variables was controlled for by logistic regression.
Table 1. Effects of Clinical Findings on the Likelihood of a Subsequent Diagnosis of Bronchitis Clinical Finding Relative Risk 95% Confidence Interval Rales 2.93 2.57-3.34 Parental expectation 2.04 1.76-2.35 that an antibiotic would be prescribed Wheezes on 1.86 1.60-2.18 examination History of sputum 1.58 1.34-1.85 Enrollment in day care 1.24 1.04-1.48 History of fever 1.22 1.04-1.42
Of the other diagnoses studied, physician perception of a parental expectation for an antibiotic prescription was associated only with viral URTIs, reducing the likelihood of that diagnosis by one half (Table 2). Diagnoses of bacterial URTIs, including otitis media and sinusitis, were unaffected by parental expectations. Parental expectation increased the probability of a diagnosis of sinusitis, made in 35 children, but the effect was not statistically significant (RR = 1.89, 95% CI, 0.9 to 4.0). Parental expectations had no effect on the physician making a diagnosis of pneumonia. [TABULAR DATA 2 OMITTED]
Parental expectation that an antibiotic prescription would be given increased the probability of this occurring (RR = 1.43, 95% CI, 1.33 to 1.55). Controlling for the other variables in the history and physical examination (fever, rales, and sputum) that were associated with a prescription for an antibiotic, the parent's expectation was second only to the finding of rales in its apparent influence in the prescribing decision.
Certain features of the child's history were associated with a parental expectation that an antibiotic would be prescribed (Table 3), but there was no difference in parental expectation due to the sex of the child (RR = 0.996), age (5.2 years among those with that expectation, 4.7 among those without), or duration of the cough (6.7 vs 6.9 days).
Table 3. Features of the Child's History That Were Associated with Parental Expectation for an Antibiotic Prescription 95% Confidence Variables Relative Risk Interval History of allergies 1.81 1.35-2.45 Enrollment in day care 1.57 1.21-2.04 Exposure to cigarette 1.32 1.02-1.70 smoke at home
Whether the clinician sensed an expectation for an antibiotic had no effect on whether a chest radiograph was done. Likewise, there was no association between the perception of a parental desire that an antibiotic be prescribed and the timing of clinical follow-up, either among children diagnosed as having bronchitis or those treated with an antibiotic.
To determine if these findings were consistent among network practice sites, separate analyses were carried out for the 14 practices that contributed 30 or more patients (1113 patients, 340 with bronchitis). When each site was analyzed separately, a statistically significant association between a parental expectation that an antibiotic would be prescribed and the diagnosis of bronchitis was found in only two sites, largely because of small sample sizes in the other 12 sites. However, with these two sites excluded and considering only the other 12, there was still a statistically significant association between a parental expectation and the diagnosis of bronchitis (P < .001 by [chi.sup.2]), suggesting that the association was not limited to a few practice sites.
The findings of this prospective network study are consistent with our hypothesis that if a clinician perceives that the parents of a child with a cough expect the physician to prescribe an antibiotic, a diagnosis of bronchitis and a prescription for an antibiotic are more likely. The association persisted even when controlling for other variables. Its magnitude was clinically significant; for example, its influence appeared to be more substantial than a history of sputum production or fever.
There are, however, three other possible explanations for the association between parental expectation and a diagnosis of bronchitis. First, it is possible that the physician diagnosed bronchitis and then the parent expressed a desire for an antibiotic. However, the perception of a parental expectation that an antibiotic would be prescribed was not associated with other diagnoses commonly treated with antibiotics, such as otitis media, sinusitis, and pneumonia. This finding suggests that there was something unique about the process of diagnosing bronchitis in these children.
Second, the perception of parental expectation that an antibiotic would be given may have been a proxy indicator for some symptom or sign within the child's illness. Although clinicians were asked to note the presence or absence of several key variables (such as fever and rales), other variables were omitted (such as the sound of the cough) that might have prompted parents to expect that an antibiotic would be prescribed and that physicians might have used in making clinical judgments. However, it seems unlikely that such a variable--something relatively obscure, but intrinsic to the disease--would would be more strongly associated with the diagnosis of bronchitis than any other clinical finding except rales.
Third, the perception of a parental expectation may have been a proxy indicator for something else, not within the child or from the parent, but coming from within the physician. The clinician may have diagnosed bronchitis and prescribed an antibiotic, and then retrospectively and incorrectly sensed that the parents expected an antibiotic. A number of such physician variables could be postulated, such as concerns about the severity of the childs illness. But in this case, one would expect to see some association between a perception of parental expectation and the scheduling of a follow-up visit or obtaining a chest radiograph. Neither association was found. Furthermore, assuming that some physician variable explains the association found in this study raises two other issues: why such a purported variable should affect only the diagnoses of bronchitis and viral URTI, and why it would be projected onto the parent rather than onto the child's disease. It seems more reasonable to make a simpler assumption, that physicians were accurately discerning (and responding to) perceptions of parental expectations. But the possibility remains that some factor that was not measured in this study is associated with both a diagnosis of bronchitis and a prescription for an antibiotic, on the one hand, and, on the other, a physician perception of a parental expectation that an antibiotic will be given.
Our findings suggest that there is a group of children whose diagnosis is unclear and for whom treatment with antibiotics is of uncertain value. Perhaps for these children, if the parents expect that an antibiotic will be prescribed, the physician may be more inclined to make a diagnosis of bronchitis and prescribe an antibiotic. Given a similar situation, in which the diagnosis is uncertain but the parent does not convey an expectation that an antibiotic will be prescribed, the physician may be more inclined to make a diagnosis of a viral URTI and not to prescribe an antibiotic. It may be within this gray zone of clinical decision making that parental expectations are more likely to influence the physician's diagnosis and treatment.
Whether the association between parental expectations and medical decisions represents a desirable or undesirable medical practice cannot be determined by this study. Are physicians simply acquiescing to parental expectations, allowing their clinical judgment to be subverted? Or are they allowing parents an appropriate role in clinical decision making? More data, both qualitative and quantitative, are needed to address these questions. As Delbanco has recently suggested, there are many potential benefits of more systematically incorporating the patient's perspective into clinical encounters.
Are there other illnesses in which physicians' decision making is affected by the patient's or parents' expectations? The findings in this study are intriguing, and may shed fight on ordinary, everyday medical decision making and on ways in which physicians let patients subtly guide that decision making.
Ambulatory Sentinel Practice Network (ASPN) Participating Practices
Alberta: Foothills Family Medicine Center, Black Diamond. Ontawo: Donald McLean, MD, Hamilton; Shelley Metcalfe, MD, and Christine Dowdell, MD, Steve Nantes, MD, Kitchener. Quebec: John Wooton, MD, Keith MacLellan, MD, and Maurice Lamarche, MD, Shawville; Centre de Medecine Familiale de Wakefield, Ltee, Wakefield.
California: Foothills Family Medical Group, Auburn; Kaiser Permanente, San Diego. Colorado: Arlis Adolf, MD, Marny Eulberg, MD, Denver; Northern Colorado Family Medicine, Greeley. Florida: James Andersen, MD, Fort Lauderdale; Family Medicine Associates, Miami. Louisiana: Linda Stewart, MD, Baton Rouge. Georgia: Titus Taube, MD. Maine: The Family Practice Center, Bangor. Massachusetts: Fitchburg Family Practice Residency Program, Fitchburg; Peter Barker, MD, Swampscott. Michigan: Medical School Family Health Center, Escanaba. Minnesota: Eagle Medical, Excelsior. New Hampshire: David Beaufait, MD, and Mark Parker, MD, Enfield; Richard Douglass, MD, Hillsboro; Manchester Family Health Center, Manchester; Monroe Clinic, Monroe; New London Medical Center, New London. New Jersey: A. John Orzano, MD, Flemington. New York: Raj Kachoric, MD. North Carolina: Aurora Medical Center, Aurora; Roanoke Amaranth Community Health Group, Inc, Jackson. North Dakota: University of North Dakota Family Practice Center, Minot. Oklahoma: Enid Family Medicine Clinic, Enid. Oregon: Dunes Family Health Care, Inc, Reedsport. Pennsylvania: Michael E. Leonard, MD, Hallstead; Highland Physicians, Ltd, Honesdale. South Carolina: Michael Zeager, MD, Taylors. Tennessee: Michael Hartsell, MD, Greeneville; James Burdette, MD, and Jan Hahn, MD, Lenoir City. Texas: Decatur Family Clinic, Decatur; Myers & Caplan Family Medicine, Mansfield; Isaac Kleinman, MD, Rosenberg. Vermont: The Health Center, Plainfield. Virginia: Edward M. Friedler, MD, Annandale; John Philip Sherrod, MD, Beaverdam; Lynchburg Family Practice Residency, Lynchburg; June Tunstall, MD, Surry; James Ledwith, MD, Tappahannock; Duane Lawrence, MD, Virginia Beach. Washington: Cle Elum Family Medicine Center, Cle Elum. West Virginia: Eglon Clinic, Eglon; New River Family Health Center, Scarbro. Wisconsin: Deforest Area Medical Clinic, DeForest; Kronenwetter Clinic, Mosinee; Terry Hankey, MD, Waupaca.
Board of Directors
Glen Aukerman, MD, Lorne A. Becker, MD, William Coleman, MD, Eugene S. Farley, Jr, MD, William L. Freeman, MD, Jack Froom, MD, Larry A. Green, MD, Curtis Hames, Sr, MD, John Feckner, MD, William Hogg, MD, Jack Kirk, MD, Walt Rosser, MD, Milton H. Seifert, Jr, MD, Richard Wasserman, MD, Maurice Wood, MD.
[1.] Pauker AG, Kassirer JP. Decision analysis. N Engl J Med 1987; 316:250-8. [2.] Balint M. The doctor, the patient and his illness. London: Pitman Medical, 1964. [3.] Heavey A. Learning to talk with patients. Br J Hosp Med 1988; 39:433-9. [4.] Squier RW. A model of empathic understanding and adherence to treatment regimens in practitioner-patient relationships. Soc Sci Med 1990; 30:325-39. [5.] Delbanco TL. Enriching the doctor-patient relationship by inviting the patient's perspective. Ann Intern Med 1992; 116:414-8. [6.] Press I. The predisposition to file claims: the patient's perspective. Law Med Health Care 1984; 12:53-62. [7.] Vinson DC. Acute bronchitis in children: building a clinical definition. Fam Pract Res J 1991; 11:75-81. [8.] Loughlin GM. Bronchitis. In: Chernick V, Kendig EL. Kendig's disorders of the respiratory tract in children. Philadelphia: WB Saunders, 1990:349-59. [9.] Stem RC. Acute bronchitis. In: Behrman RE, Kliegman PM, Nelson WE, Vaughan VC. Nelson textbook of pediatrics. 14th ed. Philadelphia: WB Saunders, 1992:1073. [10.] Denny FW, Clyde WA. Acute lower respiratory tract infections in nonhospitalized children. J Pediatr 1986; 108:635-46. [11.] Glezen WP, Denny FW. Epidemiology of acute lower respiratory disease in children. N Engl J Med 1973; 288:498-505. [12.] Cohen GJ. Management of infections of the lower respiratory tract in children. Pediatr Infect Dis J 1987; 6:317-23. [13.] Medias IB. Bronchitis in young children. Tidsskr Nor Laegeforen 1992; 112:623-4. [14.] Verheij JM, Kermans J, Kaptein AA, Wijkel D, Mulder JD. Acute bronchitis: general practitioners' views regarding diagnosis and treatment. Fam Pract 1990; 7:175-80. [15.] Green LA, Wood M, Becker L, Farley ES, Freeman WL, Froom J, et al. The Ambulatory Sentinel Practice Network: purpose, methods, and policies. J Fam Pract 1984; 18:275-80.
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||Ambulatory Sentinel Practice Network|
|Author:||Vinson, Daniel C.; Lutz, Lawrence J.|
|Publication:||Journal of Family Practice|
|Date:||Jul 1, 1993|
|Previous Article:||Interspecialty wars over endoscopy.|
|Next Article:||Extraordinary daytime urinary frequency in children.|