Use of survey research methods to study clinical decision making: referral to physical therapy of children with cerebral palsy.

Use of Survey Research Methods to Study Clinical Decision Making: Referral to Physical Therapy of Children with Cerebral Palsy Because of the lack of a strong body of research literature to support the efficacy of physical therapy in the management of physical disability, therapists, physicians, and others must operate in the realm of uncertainty when deciding whom to refer and whom to tret in a cost-effective manner. Despite the increasing passage of direct-access legislation allowing physical therapists more independence in the decision of whom to treat, physical therapy decision making remains strongly influenced by physicians' decisions regarding patient referral. This article describes a survey research approach to studying physicians' decisions regarding referral. A study, now in the pilot stage, regarding referral of young children with cerebral palsy (CP) to physical therapy will be used as an example.

Why are physicians' decisions regarding referral of children with CP important to study? First, understanding these decisions is important in a theoretical sense because we do not know what factors guide physicians' referral decisions and because the research literature on pediatric physical therapy is unclear regarding what the expected outcomes of treatment should be. Outcomes may differ by type of CP and by severity. The lack of data describing such possible differences in outcome makes it impossible to use quantitative decision analysis approaches to selecting management options, such as intensive physical therapy, orthopedic surgery or neurosurgery, or use of rehabilitation technology; for this population. Studying the opinions of decision makers of different levels of expertise and experience can help to clarify the theory, reveal where the opinions of physicians diverge from those of physical therapists, and lend direction to future research.

Second, physicians appear to be increasingly aware of the lack of scientific support for the efficacy of physical therapy in treating CP and of the high cost involved in the long-term care of children with developmental disabilities. In an era when third-party payment for such services is threatened, lack of accountabiliity looms large as a problem. To give just one example, in an article on the economics of long-term care of children following neonatal intensive care, the authors indicated that physical therapy and occupational therapy were responsible for the greatest costs in the area of outpatient services following hospital discharge, and those children at intermediate risk for developmental disability used these resources as much as children at exceptionally high risk. [1] The authors also noted the lack of documented efficacy of these services. An understanding of how physicians deal with the uncertainty involved in early diagnosis of a condition with undocumented benefit from therapy can be helpful in developing marketing strategies for increasing referrals while we await the needed research.

Third, study of expected positive and negative outcomes of therapeutic intervention can aid in understanding how physicians deal with conflicting goals, in thsi case, that of promoting high-quality parent-infant interaction during infancy and that of preventing the development of contractures and other possible disabling outcomes of motor dysfunction.

The aims of our study, then, are 1) descriptive (ie, to describe physicians' general propensity to refer children with CP to physical therapy and the nature of their belief in the value of physical therapy to ameliorate the motor dysfunction of such children; 2) to study how referral is affected by the experience and training of the physician, the symptomatology of the children, and the physicians' opinions regarding efficacy of physical therapy; and 3) to study how systematic and reliable physicians' decisions are when faced with actual cases in which they must decide, first, whether the child has CP, and, second, whether the child should or should not be referred to physical therapy. These aims will be accomplished through survey research in which physicians will be asked to make decisions regarding the disposition of several specific cases involving children with motor dysfunction. After reading each brief case history, the physicians will review a short videotape showing the child's movement and then answer a series of questions regarding their diagnosis, referral decision, and expected outcomes of therapy.

Review of the Literature

Uncertainty in Clinical

Decision Making

On the surface, the referral decision seems simple--a physician identifies a child as having CP and refers the child for evaluation and treatment. On deeper investigation, however, such decisions are fraught with uncertainty, the very area in which systematic clinical decision-making approaches are most fruitfully applied. To explicate further, first, the physician must accurately diagnose the problem, that is, recognize that the child has a motor problem (eg, delayed motor development, abnormal neurologic signs, symptoms that the parent believes to mean that something is wrong). Parents of children with CP frequently state they believe their physical denied that a problem existed or failed to recognize it. Upon review of the literature, however, one finds that much controversy exists among medical professionals regarding risk factors for CP, how to classify the motor involvement characteristic of the rather heterogeneous entity consisting of the cerebral palsies, the earliest symptoms, and how early a definitive diagnosis can be made. [2] These uncertainties are superimposed on a field in which the causes of CP have been changing while no definitive natural history of the disorder is available. Furthermore, research has revealed that the most commonly used developmental examination to identify motor delay is out-of-date and no "good-standard" diagnostic tool is readily available. [3]

Efficacy of Physical Therapy

Once a diagnosis has been made, the physician must decide that physical therapy has potential value in order to make a referral. In support of the physician's decision, we find that the research literature offers studies mostly on only one form of physical therapy, neurodevelopmental therapy (NDT), and is focused on whether the rate of motor development can be improved. [4] Neurodevelopmental theory, however, suggests that posture, coordination, and balance are the outcome variables affected by NDT. Thus, it appears that either physical therapists have been imprecise in specifying expected outcomes of treatment or researchers have misinterpreted therapy claims. That aside, the literature to date does not demonstrate high efficacy. Ottenbacher and colleagues, for example, performed a meta-analysis of the studies of NDT efficacy completed prior to 1985 and concluded that about 62% of children receiving NDT did better than control subjects (about 50% would be expected to do better by chance alone). [4] The most recent study, and one of the best-designed to date, by Palmer and colleagues, suggests that children with spastic diplegia will achieve motor milestones faster with a developmental stimulation program than with physical therapy. [5]

Under conditions of uncertainty regarding diagnosis and efficacy of treatment, it seems appropriate to study whether physicians will refer a child identified as having motor dysfunction in infancy and what specific outcomes they might anticipate. Because previous research on medical decision making has shown that positive outcomes to be expected are not the only determinants of treatment choices, the referral decision can be expected to be influenced by a number of different factors. [6] Physicians have been shown previously to be risk aversive and will sometimes choose no treatment, even when research demonstrates high efficacy if a risk of harm is also involved in the decision to treat. Thus, various factors including the diagnosis and its certainty; the severity of the dysfunction; the parents' wishes; and the physician's training, specialty, and experience are all possible influences on the decision to refer to physical therapy.

Physicians' Attitudes Toward

Early Intervention 

Physical therapists and parents of children with CP often complain that referral by physicians is delayed too long. Development occurs at such a rapid pace in the first two years of life that delayed referral means loss of crucial time at the most formative stage of a child's motor development. Furthermore, delayed referral favors formation of bad movement habits that may be difficult to alter and that may lead to development of deformities. The research literature indicates that physicians may delay referral for a number of reasons, including 1) the uncertainty of early diagnosis of CP, 2) the belief that intervention is not efficacious or is not necessary, and 3) the concern that early involvement in efforts to promote motor development will interfere with establishment of parent-child bonding and place undue demands on parents to be therapists rather than parents. [6,7] Overall, although surveys of physicians' attitudes toward early intervention suggest that they believe in a modest impact on children and substantial positive effects on family members, they demonstrate physicians' reluctance to diagnose and label developmental problems and that they appear to refer children with chronic conditions to specialists less than is desirable. [8-12]

Previous research on physicians' attitudes toward referral of children with developmental problems to early intervention specialists is limited primarily to information on the opinions of pediatricians regarding early intervention. For example, in a survey involving 354 pediatricians, Guralnick and colleagues found that children with mild CP (spastic diplegia) were ranked as fourth most likely to benefit directly from early intervention among a list of children with various types of developmental disabilities. [8] Severely involved children were not believed to receive significant individual benefit. In cases of motor dysfunction, the effect of intervention was likely to be judged as potentially greater for the family than for the child. Overall, the physicians were positive in believing in a modest impact on children and substantial effects on the family. About 50 pediatricians noted actual contraindications to enrollment in early intervention programs, including raising false expectations, financial stress, and concerns about lack of demonstrated efficacy. Similarly, Esposito found that, among a small sample of pediatricians and pediatric neurologists (N = 28), the majority believed early intervention to be beneficial for the infant, and all of the physicians believed that benefits accrued to the family members. [7] The physicians interviewed by Esposito differed in their opinions regarding the specific effects of intervention on the child. Pediatricians expressed significantly greater belief than neurologists in the efficacy of early intervention in promoting motor, cognitive, and social development. On the other hand, prevention of contracture and deformities was endorsed as a likely benefit by almost all of the physicians. Responses in this study were not elicited to specific cases, so it is not known whether these physicians would have anticipated varying outcomes for children with different disabilities. Neither of these studies elicited opinions regarding the specific effects of physical therapy.

Underreferral of Children with

Developmental Disabilities

The only previous study in the area of intervention expectations with developmentally disabled children that used a design similar to ours is a study of Goodman and Cecil. [9] The results of their study indicated that pediatricians would have referred for developmental programing far fewer children with mental retardation than was warranted by the results of research studies on efficacy of programs to promote cognitive development in similar children.

Other work has also demonstrated the extent of unmet health care needs in children with long-term disabilities. [10-12] In a study of children with spina bifida, for example, only 18% of children had comprehensive coverage of their health care needs. [10] Only 9% of primary care physicians in this study provided families with significant help in management of the chronic condition. Pless and colleagues specifically examined the use of allied health services, including physical therapy, by pediatricians and general practitioners in upstate New York. [11] Significant variation among physicians in the use of services for children with CP was noted. Some physicians always referred children with CP to physical therapy, whereas others never did. About 40% of pediatricians referred children with CP to physical therapy "always" or "often." The differences could not be explained by specialty, by availability of services, or by frequency of referral to medical specialists. The authors suggested that failure to use allied health services might be related to beliefs that these services were of little value; however, they did not collect data regarding attitudes toward efficacy. They also did not compare referral patterns of primary care physicians with those of medical specialists in pediatric rehabilitation. Nonetheless, the authors suggested that allied health services are underutilized by primary care physicians and that serious deficiencies exist in both medical education and the health care system that result in less than optimal care for physically disabled children.

Compliance with Referral

Recommendations

Another study relevant to this topic is a survey of compliance with consultants' recommendations. Only about 73% of 143 recommendations to 77 clients (families, centers, professionals) working with 30 developmentally retarded children were carried out. [13] Compliance did not differ by child or client factors. Of those factors contributing to variance in compliance, completeness of explanation of the rationale for the treatment recommendation was the strongest predictor. Feasibility of carrying out the recommendation was the second most important. Last, but also predictive, was the recipient's belief in the efficacy of what was recommended, suggesting that this factor will be likely to be important in the decision to refer as well. Compliance with rehabilitation programs, in general, ranges from 34% to 67%, [14] so the reported compliance in this study was actually quite high. The authors concluded that attitudes and beliefs are most important in determining willingness to carry out the recommendations of consultants regarding decisions about treatment or other special programs for handicapped children. Of special note is that the article ends with a caveat that consultants should be cautious about implementing programs to increase compliance when the efficacy of various interventions has not been demonstrated.

Diagnostic Uncertainty

A further problem that physicians face in deciding whether to refer an infant with suspected motor dysfunction to physical therapy is the uncertainty of early diagnosis. Of particular concern to this investigation is a commonly held assumption that CP cannot be diagnosed in the first year of life. This assumption has been based on a small amount of data on children who "outgrew" CP. [15] The article appears to have had a sizeable impact on referral decisions, especially in the first year of a child's life. The assumption that follows is that children should not be referred for treatment even if they clearly have motor deficits. What is very unclear is when the problem warrants referral and whether the published data are accurate in generating the assumption that diagnosis is unreliable in the first year.

The article that led to concern regarding early diagnosis presented data from a prospective study of 50,000 pregnancies in which 118 of 229 children with a diagnosis of CP at the age of 1 year were no longer believed to have it at age 7 years. [15] The validity of these data is seldom challenged, but it is important to note that, although these children did not retain the CP diagnosis, a large proportion of them were not "normal" at age 7 years. Twenty-two percent had mental retardation, 19% were hyperactive, 21% were immature behaviorally, and some had other less serious handicaps, such as visual disorders. In all, 40% had one or more problems of a sensory or cognitive nature, many of which would warrant referral to physical therapy or occupational therapy. The majority of children whose CP resolved by age 7 years were considered to have mild problems at 1 year of age; the majority of children with moderate or severe CP at age 1 year retained the diagnosis. Two other published reports are pertinent here. In one from a distinguished developmental diagnostic clinic, retrospective review of records on children with CP showed no false positive results. n16 In the second report, a retrospective Danish study covering 30 years and 2,100 diagnoses, the authors found only 17 patients for whom development truly appeared to have normalized. [17]

The Pilot Study

Conceptual and Analytic

Framework

Before moving on to provide more of the specific details of this study, it should be placed within a theoretical framework. The theoretical bases of this study lie in two areas of research on clinical decision making--clinical judgment theory and, to some extent, decision analysis theorey, both quantitative statistical approaches to the study of treatment or diagnostic decisions. The first, clinical judgment theory, uses multiple regression statistics to describe the clinicians' decisions, in this study, whether a child has CP and whether to refer that child to physical therapy. [18-21] Clinical judgment theory suggests that clinical decisions can be modeled by equations involving simple combinations of variables known as cues. In this study, we plan to ask physicians what clinical signs were used to decide that a child has CP or another diagnosis, along with the probabilities they attach to each possible diagnosis. Multiple regression equations will also be used to describe the judgment to refer, based on the characteristics of the physician (eg, experience, training, specialty), the characteristics of the child (eg, diagnosis, severity of the disorder), and the attitudes and beliefs of the physician regarding the value of physical therapy and his or her general propensity to refer children with CP to physical therapy.

A second analysis we will undertake, that of the relationships among physician beliefs in the value of various possible outcomes if therapy is prescribed relative to the likelihood of their occurrence, is related to decision analysis theory. [6,22-26] Assessments of expected outcomes and their probability of occurrence can be used for modeling prescriptions and providing advice to physicians (and physical therapists interested in improved marketing of their services) on ways to improve clinical decisions in the face of uncertainty. Such work is of great importance, because research has shown that physicians often make decisions that do not necessarily reflect their own beliefs in the utility of various choices.

Elstein and colleagues, for example, studied physicians' decisions regarding the use of estrogen replacement thereapy in postmenopausal women. [6] They found that both the research literature and the physicians' own descriptions of the efficacy of such treatment in relieving symptoms and preventing fractures secondary to osteoporosis versus the relative risk of causing endometrial cancer dictated that estrogen treatment should always be used, even in women at high risk for cancer. Yet, only about one third of the physicians' decisions were to prescribe treatment. The results suggested that physicians found the risk of cancer, however remote, to be so aversive that they chose to ignore the present value of relieving menopausal symptoms and the far more important future value of preventing highly likely painful and disabling fractures that are so common in elderly women. Although research results may take a long time to filter down into clinical practice, the researchers were able to show that physicians did indeed know the probability of these risks and outcomes well but behaved in a way that disagreed with their knowledge because they were seemingly risk aversive and preferred to avoid doing harm actively (ie, risking causing cancer) in favor of just letting things happen (ie, failing to prevent osteoporosis and resultant fractures).

In our work, of course, we are not avoiding risk of death, but failure to provide early physical therapy for children with CP certainly could be conceived as allowing disability in the form of contractures and expensive resulting surgery to occur. The alternative decision--to refer--involves questions of efficacy and large costs. Much more uncertainty exists regarding the efficacy of treatment, and it is unknown how well physicians know the literature on treatment efficacy because much of it is not published in medical journals. We also need to know how strongly beliefs in avoiding labeling, the uncertainty of early diagnosis, and the risk of negatively affecting parent-infant interaction will lead to risk-aversive behavior rather than emphasizing the possibility of improving motor development, preventing contractures and surgery, and improving the family's ability to cope with handicapped child, suggested by Cadman and colleagues to be important factors in physician belief systems. $(13$)

Methodology

As indicated earlier, the method we will use to explore these questions is survey research involving a questionnaire to be completed by physicians. The sample (N = 200) will consist of members of the American Academy for Cerebral Palsy and Developmental Medicine (a specialist group of physicians) and members of the American Academy of Pediatrics (representing primary care physicians as well as pediatric specialists.) The questionnaire consists of questions relating to the physicians' training, experience, practice, and general beliefs in the efficacy of physical therapy for children with CP followed by presentation of a series of cases of children under the age of 2 years with suspected motor dysfunction. The physicians will read a brief case history and then watch a videotaped vignette showing the child's motor skills.

As an example of a typical case, the following is one child's description:

Baby John was born to a 26-year-old mother at 43 weeks gestational age, weighed 3,650 g, and had a five-minute Apgar score of 9. During the neonatal period, he suffered a right middle cerebral artery infarct, apnea, bradycardia, and seizures confirmed by electro-encephalography. John is currently 18 months old with head circumference at the 5th percentile, length at the 80th percentile, and weight at the 8th percentile. On a standardized test of gross motor development, he received a developmental quotient of 70.

Following exposure to each case, the physician will be asked to decide whether the child is normal, has transient dystonia, or has CP and what the probability of each diagnosis is (progressive neurologic conditions have been ruled out to keep the problem simple). The physician will be asked to state what signs and symptoms entered into his or her decision and whether the child would be referred to physical therapy. If the decision to refer is made, the physician will be asked to state what the possible expected outcomes are and how likely they are. Outcomes include the possibility of motor performance improvement (eg, improved muscle strength, endurance, coordination), prevention or correction of contractures, and improvement in cognitive development and family functioning. The cases to be presented range from normal children through the high-risk children to those with severe handicaps so that the efficacy of treatment can be studied for different types of children.

We hypothesize that

1. Physicians, in general, wil demonstrate low frequencies of referral of the children with CP in this study.

2. Physicians will be more likely to refer children with CP to physical therapy if the child is older, the dysfunction is severe, and there is certainty of diagnosis.

3. Physicians will be more likely to refer children with CP if they have training focused on pediatric disabilities, a general propensity to use other medical services, and high expectations of physical therapy efficacy.

4. The major reasons for referral of the child with CP to physical therapy will be related to prevention of contractures and enhancement of the family's ability to cope with the child.

5. The major reasons for not referring the child with CP to physical therapy will be related to the negative effects of labeling and to cost.

Referral decisions will be analyzed with multiple regression equations to explain the variance in referral decisions based on physician and child characteristics. Descriptive presentations of the physicians' beliefs in the efficacy of physical therapy will also be developed.

Conclusion

It is hoped that this description of the theory and procedure underlying a planned study of clinical decision making involving referral to physical therapy will spark imitators. Survey research involving both physicians and physical therapists and regarding how they make decisions under the uncertainty of diagnostic unreliability and lack of a solid and extensive research base supporting intervention can provide important data in generating new theory, testing long-held assumptions, and clarifying the bases of our practice.

References

[1] Shankaran, S, Cohen SN, Linver M, et al: Medical care costs of high-risk infants after neonatal intensive care: A controlled study. Pediatrics 81:372-378, 1988

[2] Rosenbloom L, Thomson APJ: Cerebral palsy: A 1987 perspective. Pediatrician, to be published

[3] Campbell SK, Siegel E, Parr CA, et al: Evidence for the need to renorm the Bayley Scales of Infant Development based on the performance of a population-based sample of twelve-month-old infants. Topics in Early Childhood Special Education 6(2):83-96, 1986

[4] Ottenbacher KJ, Biocca Z, DeCremer G, et al: Quantitative analysis of the effectiveness of pediatric therapy: Emphasis on the neurodevelopmental treatment approach, Phys Ther 66: 1095-1101, 1986

[5] Palmer FB, Shapiro BK, Wachtel RC, et al: The effects of physical therapy on cerebral palsy: A controlled trial in infants with spastic diplegia: N Engl J Med 318:803-808, 1988

[6] Elstein AS, Holzman GB, Ravitch MM, et al: Comparison of physicians' decisions regarding estrogen replacement therapy for menopausal women and decisions derived from a decision analytic model. In Dowie J, Elstein A (eds): Professional Judgment: A Reader in Clinical Decision Making. New York, NY, Cambridge University Press, 1988, pp 298-322

[7] Esposito RR: Physicians' attitudes toward early intervention. Phys Ther 58:160-167, 1978

[8] Guralnick MJ, Heiser KE, Eaton AP, et al: Pediatricians' perceptions of the effectiveness of early intervention for at-risk and handicapped children. J Dev Behav Pediatr 9(1):12-18, 1988

[9] Goodman JF, Cecil HS: Referral practices and attitudes of pediatricians toward young mentally retarded children. J Dev Behav Pediatr 8(2):97-105, 1987

[10] kanthor H, Pless B, Satterwhite B, et al: Areas of responsibility in the health care of multiply handicapped children. Pediatrics 54: 779-785, 1974

[11] Pless IB, Satterwhite B, Van Vechten D: Chronic illness in childhood: A regional survey of care. Pediatrics 58:37-46, 1976

[12] Stein REK, Jessop DJ, Riessman CK: Health care services received by children with chronic illness. Am J Dis Child 137:225-230, 1983

[13] Cadman D, Shurvell B, Davies P, et al: Compliance in the community with consultants' recommendations for developmentally handicapped children. Dev Med Child Neurol 26:40-46, 1984

[14] Sackett DL, Snow JC: The magnitude of compliance and noncompliance. In Haynes RB, et al (eds): Compliance in Health Care. Baltimore, MD, The Johns Hopkins University Press, 1979, pp 11-22

[15] Nelson KB, Ellenberg JH: Children who "outgrew" cerebral palsy. Pediatrics 69:529-536, 1982

[16] Farber JM, Shapiro BK, Palmer FB, et al: The diagnostic value of the neurodevelopmental examination. Clin Pediatr (Phila) 24:367-372, 1985

[17] Taudorf K, Hansen FJ, Melchior JC: Spontaneous remission of cerebral palsy. Neuropediatrics 17:19-22, 1986

[18] Elstein AS, Bordage G: Psychology of clinical reasoning. In Dowie J. Elstein AS (eds): Professional Judgment: A Reader in Clinical Decision Making. New York, NY, Cambridge Univesity Press, 1988, pp 109-129

[19] Elstein AS, Shulman LS, Sprafka SA: Medical Problem Solving: An Analysis of Clinical Reasoning. Cambridge, MA, Harvard University Press, 1978

[20] Elstein AS, Holmes MM, Ravitch MM, et al: Medical decisions in perspective: Applied research in cognitive psychology. Perspect Biol Med 26:486-501, 1983

[21] Johnson PE, Hassebrock F, Duran AS, et al: Multimethod study of clinical judgment. Organizational Behavior and Human Performance 30:201-230, 1982

[22] Fischer AF, Stevenson DK: The consequences of uncertainty: An empirical approach to medical decision making in neonatal intensive care. JAMA 258:1929-1931, 1987

[23] Pauker SG: Coronary artery surgery: The used of decision analysis. Ann Intern Med 85:8-18, 1976

[24] Sackett DL, Haynes RB, Tugwell P: Clinical Epidemiology: A Basic Science for Clinical Medicine. Boston, MA, Little, Brown & Co Inc, 1985

[25] Slovic P. Lichtenstein S: Comparison of Bayesian and regression approaches to the study of information processing in judgment. Organizational Behavior and Human Performance 6:649-744, 1971

[26] Shewchuk RM Francis KT: Principles of clinical decision making: An introduction to decision analysis. Phys Ther 68:357-359, 1988

S Campbell, PhD, FAPTA, is Professor, Department of Physical Therapy, College of Associated Health Professions, University of Illinois at Chicago, 1919 W Taylor St, Chicago, IL 60612 (USA).

J Anderson, MS, PT, is Graduate Research Assistant, Department of Physical Therapy, College of Associated Health Professions, and Doctoral Student, Department of Sociology, University of Illinois at Chicago.

HG Gardner, MD, is Pediatrician, DuPage Pediatrics, Ltd, 805 Plainfield Rd, Darien, IL 60559, and President, Illinois Academy of Pediatrics.

This work was supported in part by Grant MCJ 9101, Bureau of Health Care Delivery and Assistance, US Public Health Service, to the Department of Occupational Therapy, University of Illinois at Chicago, and by intramural research funds from the College of Associated Health Professions, University of Illinois at Chicago.