< Back to Articles Index             

Violence and Victims, Vol. 1, No. 2, 1986

Consensus and Difference Among Hospital Professionals in Evaluating Child Maltreatment

Jane C. Snyder* Eli H. Newberger*

The decision-making process in suspected cases of child maltreatment involves reaching interprofessional consensus. Interprofessional consensus in seriousness ratings of maltreatment incidents for the welfare of the child was examined by surveying 39 case vignette ratings by 295 pediatric hospital professionals from five occupations. The survey instrument was derived from research by Giovannoni and Becerrra (1979). An exploratory factor analysis yielded five categories of maltreatment: physical abuse, sexual abuse, general failures in care, minor neglect/discipline, and lifestyles/values. A sixth category, parental sexual preference, was rated not very serious and did not appear to belong in the maltreatment domain. Nurses and social workers rated incidents as most serious, differing significantly from psychiatrists and, often, from physicians and psychologists. Professions agreed on rank ordering of categories by seriousness. Variables such as sex, parenthood status, years of experience, and medical specialty showed some relationship to ratings within some professional groups.

Medical and mental health professionals and educators are mandated by law to identify and report cases of suspected child maltreatment. Identifying such cases is not an easy task. A number of clinicians have written on the emotional and situational concomitants of identification and reporting (e.g, Elmer, 1960; Helfer, 1975; Hill, 1975; Rosenfeld & Newberger, 1977; Rosenzweig, 1982). Definitions of child abuse in state reporting statutes list general areas of concern such as physical abuse, physical neglect, sexual abuse, emotional abuse, and educational and medical neglect. The vague nature of these definitions contributes to difficulty in case identification. Defining specific constituents of abuse or neglect is left to the professional. Filing a report of suspected maltreatment initiates an investigation by the state's protective services division, resulting in further definition; the case is either "substantiated" or closed. State agency guidelines regarding which cases will be substantiated are as likely to change with budget considerations as with attention to conditions adversely affecting children (Newberger et al., 1975).

Many have argued that applying the label of child abuse or neglect implies judgment about social deviance and brings to bear both personal and societal values regarding parenting (e.g., Gelles, 1973; Giovannoni & Becerra, 1979; Nagi, 1977). Implicit in the process of defining child abuse are judgments about circumstances harmful to children, minimal requirements for adequate child development, aspects of a child or person that are most important for society to protect. Such judgments may be affected by a number of subjective variables, including: personal experience as a child and parent (Kaufman, 1983), social group affiliation accompanying norms for appropriate child care (Daniel, 1985; Korbin, 1977), personal values (Rosenzweig, 1985). It is also likely that the professional training has received will affect the standard used in defining maltreatment, as diffe occupational groups have different roles in working with children and families, concern themselves with different aspects of child development.

Several research studies have investigated differences among professional groups in assigning the label of child abuse to case vignettes (Gelles, 1975; Nalepka, O'Toole, & Turbett, 1981; Turbett & O'Toole, 1980). Differences among professionals in these studies were most apparent on vignettes which presented ambiguous intentionality of action on the part of the caretaker, and in cases where effects on the child were not obvious. High agreement occurred on "the outrageous cases," e.g., cigarette burns (Gelles, 1975). However, these studies and others (e.g., Hampton & Newberger, 1985; Katz et al., 1986) have documented that factors such as race and social class are as important in determining which cases will be labeled as child maltreatment and in affecting the disposition of those cases, as is the nature the injury or incident, particularly given ambiguous circumstances.

Many pediatric hospitals employ interdisciplinary teams to facilitate the decision-making process in suspected cases and to guard against the subjective of anyone individual or professional viewpoint. If a formal team does not exist, an informal case management conference may be held. The difficulties in arriving at consensus, documented in research studies, have also been observed in clinical practice (e.g., Bourne & Newberger, 1980), where most cases are, in fact, ambiguous,1 The greater the extent of disagreement among team members, the more difficult it becomes to devise an adequate case plan.

In this study, the degree of consensus among occupational groups involved in clinical decision-making in suspected cases of child abuse was investigated in a pediatric hospital setting. The survey instrument was adapted from methodology used in a study of the child abuse definition process (Giovannoni & Becerra, 1979) in which participants were asked to assess seriousness for the welfare of the child in 78 case vignettes.

Giovannoni and Becerra studied five professional groups: social workers, police, pediatricians, lawyers, and a lay sample. Significant disagreement among professional groups occurred on all but nine vignettes. All groups tended to agree in relative ranking of categories of maltreatment by seriousness-rating physical at sexual abuse, and the fostering of delinquency as most serious. "Educational neglect and failure to provide" preceded "parental sexual mores," which were viewed as least serious. Inadequate supervision and emotional maltreatment were rated between the two extremes.

Giovannoni and Becerra used two forms, with one set of vignettes adding consequences for the child. This addition significantly increased the ratings of seriousness for over half the vignettes, decreased ratings for nine, and made no difference in 26 cases. Relative ranking according to seriousness remained the same with and without consequences, however. Medical neglect was the category in which professionals were most affected by the addition of consequences. In this category, seriousness ratings decreased for pediatricians and increased for the other groups.

Since the general findings were the same with and without consequences, and since knowledge of consequences is part of the information used in clinical decision-making in hospital settings, vignettes used in the present study included this dimension.

METHOD

Subjects

Members of five professional groups at an urban pediatric teaching hospital and an affiliated mental health clinic participated in the study: physicians (nonpsychiatrists), nurses, social workers, psychologists, and psychiatrists. Physicians, psychiatrists, and psychologists were surveyed at the start of in-service teaching sessions on child abuse or related topics at the hospital or affiliated clinic. Nurses and social workers were surveyed in two ways: 64% (N = 53) of the social workers and 36% (N = 31) of the nurses were contacted at the beginning of in-service teaching sessions. In order to obtain more participants from these groups, the questionnaire was mailed to all hospital social workers and social work students and to nursing staff on hospital divisions frequently involved with child abuse cases. A 60% response rate (N = 30) was obtained from social work staff, and 40% (N = 56) from nursing staff. (The social work sample included 14 protective social workers attending a hospital teaching conference. Their responses were analyzed separately from the rest of the social work group, but did not differ significantly and were combined into the total sample.)

Of the 306 returned questionnaires, 295 met the study criteria, i.e., were complete, and came from professionals within the five disciplines. Tables 1 and 2 summarize subject characteristics for each professional group, including sex, age, parenthood status, years of experience posttraining, and prior involvement with cases of abuse or neglect.

There is no page 128.

RESULTS

Factor Analysis

A factor structure was predicted based on the Giovannoni and Becerra nine-factor structure. The predicted structure was not confirmed by confirmatory factor analysis of these data. An exploratory factor analysis with Varimax rotation was then performed. Six factors emerged from this analysis, explaining 70% of the variance. Table 3 lists the items and factor loadings. The factors, in order of variance accounted for, were: general failures in care, parental sexual preference, sexual abuse, physical abuse, minor neglect and discipline, and lifestyles/values.

Professional Differences and Consensus

One-way analyses of variance were performed to assess professional differences in ratings on each item and on each factor. Item-by-item comparisons yielded significant differences in mean professional ratings on 27 of 39 items. A comparison of mean professional ratings on each factor using one-way analysis of variance revealed significant differences on all factors except the physical abuse category. These findings are summarized in Table 4. The predominant pattern of consensus in ratings was agreement between nurses and social workers, who rated items as significantly more serious than did pediatricians and psychiatrists. Psychologists were between these two extremes, and did not differ significantly from the other groups. The exception to this pattern occurred for factor 2-parental sexual preference-to which nurses and pediatricians gave significantly higher seriousness ratings that did social workers. With one exception, all professional groups were in agreement regarding rank ordering of seriousness of maltreatment categories. For psychiatrists, the "minor neglect/discipline" category was rated as more serious than "lifestyle/values."

Note: Principal components factor analysis with Varimax rotation. These six factors accounted for 70% of the variance. The factor "general failures in care" accounted for 50% of the variance. A factor loading of .50 or better was used as a criterion for inclusion on a factor. ^aThis item also loaded on the factor "general failures in care."      

Important Personal Variables

Sex Differences. Significant sex differences emerged on all factors except parental sexual preference, in t-test comparisons.³ Since profession and sex were interrelated, sex differences were examined separately using t-tests within each professional group that contained enough members of each sex to analyze: social workers, pediatricians, and psychologists. No significant sex differences were found for social workers. In the psychologist and physician groups, one significant difference was found in each group. Female psychologists rated seriousness of general failures in care significantly higher than males. For physicians, females rated seriousness of physical abuse significantly higher than their male colleagues (Table 5).

Parenthood Status. A comparison of mean seriousness ratings of maltreatment categories for subjects who were or were not parents was performed. This was analyzed for the sample as a whole and for each professional group, using t-tests. No significant difference according to parenthood status emerged for the sample as a whole. Within the social work and psychologist groups, however, significant differences were found for social workers in ratings of sexual abuse and lifestyle/values, and for psychologists in ratings of parental sexual preference. In cases where differences emerged, parents rated the category as significantly more serious than did non parents (Table 6).

Note: Groups sharing an underlining agree; groups differing in underlining differ significantly from each other.
^aFrom one way analyses of variance; a posteriori comparisons using Duncan's multiple range test.
^bSocial workers differed significantly from nurses and physicians on this category.
*p < .05; .**p < .01; ***p < .001.

^aFrom two-tailed t-tests.
^bFor this comparison, variances were significantly different and Satterthwaite's approximation was used.
*p < .10; **p < .05.

Years of Experience Posttraining. Correlations of years of experience by category seriousness ratings were performed for each professional group. A significant relationship was found only for nurses, occurring on all six factors. Years of experience correlated negatively with seriousness ratings on each factor (Table 7).

^aFrom two-tailed t-tests.
^bFor this comparison, variances were significantly different and Satterthwaite's approximation was used.
*p < .05.

^aPearson's Rho.
*p < .05; **p < .01; ***p < .001.

Prior Involvement With Cases of Child Abuse. This variable was not significantly related to seriousness ratings for any professional group.

Medical Specialty. A comparison of mean severity ratings of categories for three groups of physicians by medical speciality-medical students, pediatricians, and surgeons- indicated that surgeons rated seriousness lower on all categories than did the other doctors. These differences were significant for three categories; general failures in care, minor neglect, and lifestyle/values (Table 8).

DISCUSSION

Defining Child Abuse: Categories of Maltreatment

Results of this study suggest that professionals do discriminate among types of child maltreatment, and are in some consensus regarding the relative seriousness of these categories for the welfare of the child. The failure to confirm the factor structure predicted based on the Giovannoni and Becerra study findings is not surprising, given the differences between the two studies. The current study uses only half the number of vignette items (39 of78). In addition, Giovannoni and Becerra's sample included only one hospital-based group of professionals, pediatricians, whereas the present study was designed to assess seriousness ratings by medical and mental health professionals in a hospital setting.

Note: Groups sharing an underlining agree. Groups differing in under-linings differ significantly from each other.
^aOne-way analyses of variance; a posteriori comparisons between means using Duncan's Multiple Range Test.
^bp < .10; *p < .05; **p < .01.

While fewer factors emerged in this study, three of the six factors are conceptually similar to factors in the Giovannoni and Becerra analysis. These are: physical abuse, sexual abuse, and general failures in care. Some categories which emerged as separate factors in the Giovannoni and Becerra study were subsumed under the "general failures in care" factor in this study. The two weakest factors in this study - minor neglect/discipline, and lifestyle/values - did not have conceptual parallels among the Giovannoni and Becerra factors, but included items which loaded on a number of different factors in their study (e.g., failure to provide; fostering delinquency, drugs, and alcohol; sexual mores) or did not load on any factors in their study (e.g., the spanking item).

Each of these two factors in the current study included items representing a variety of caretaking situations. For the lifestyle/values factor, the similarity among constituent items is one of lifestyle or moral deviance from the cultural norm: fostering criminal behavior, use of drugs, and open sexual behavior. The exception is the spanking item, which conceptually would seem to fit with the "minor neglect/ discipline" items. Spanking, however, is controversial, being socially acceptable to some groups and unacceptable to others (e.g., Erlanger, 1975; "The Last? Resort," 1985; Straus, Gelles, & Steinmetz, 1980). Thus, use of corporal punishment may be considered a lifestyle difference, although it differs from other items found in this category.
Among the "minor neglect/discipline" items, the commonality appears to be in the domain of physical care, but with less serious consequences for the child than for items loading on "general failures in care."

The "parental sexual preference" factor in this study emerged as a conceptually distinct category, not loading with other lifestyle items. (In the Giovannoni and Becerra analysis, these items loaded with other "sexual more" items not included in this study.) This category was also rated as least serious for the welfare of the child. The mean rating for the entire sample was 3.29, a full 3 points below the next highest category and in the "not at all serious" range. These findings strongly suggest that parental sexual preference does not belong in the domain of child maltreatment.

Differences Among Professional Groups

Similar to the findings of earlier studies (Gelles, 1975; Turbett & O'Toole, 1980), consensus within this study was greatest on the severe physical abuse items, where the maltreatment appeared intentional and resulted in physical harm to the child.

The consistent patterns of agreement and disagreement among professional groups (in mean ratings of other categories of maltreatment, as well as of individual items) point to differences in training and work roles. This may reflect the fact that nurses and social workers tended to rate vignettes and categories at the high end of the seriousness scale. These two professions are the "front line" for inpatient cases in which there is a question of maltreatment. Nurses have the most frequent contact with the child, as well as ongoing contact with parents. Social workers are in contact with parents, attempting to forge a relationship and to complete an assessment. The training of both professions heavily emphasizes social and psychological factors in child and family functioning, thereby increasing sensitivity to more subtle kinds of child maltreatment and potential consequences. The observations nurses and social workers make are usually critical in decision-making about the filing of child abuse reports. Hence, the burden of recognition may fall heavily on their shoulders.

One would expect psychologists and psychiatrists to have training emphases similar to that of nurses and social workers, and to be sensitive to social and psychological circumstances as well; yet psychiatrists consistently rated incidents as less serious than any other group, with physicians next lowest, preceded by psychologists (Table 4). Psychologists and psychiatrists are involved with possible cases of child maltreatment in two ways: as diagnostic consultants on inpatient cases, and as ongoing therapists. In the former role they have fleeting contact, often a single interview with patient and parents, or perhaps no contact with parents at all. While their observations are important, their temporal and emotional involvement is limited. In their role as therapists, their focus is primarily on potential long-term effects and on preventive intervention. In the vignettes, however, consequences to the child were situational and immediate. Long-term consequences had to be inferred by the reader. It is possible that psychologists and psychiatrists, trained to ameliorate long-term psychological damage, may de-emphasize the "harm" produced from isolated incidents or parenting mishaps. Their training embodies the belief that intervention can ameliorate psychological damage and that single incidents do not in themselves produce irreversible harm.

The relatively lower seriousness ratings by physicians, who would be expected to have daily contact with hospitalized patients and to have the prime role in case management, is probably best explained by their biomedical as opposed to psychosocial orientation. This focus emphasizes physical symptomatology to the exclusion of psychological difficulties. In addition, physician involvement with the patient and family, though frequent, is more fleeting than that of nurses and social workers. Physicians' relatively low ratings compared to nurses and social workers parallels their place in the rankings by professional groups in the Giovannoni and Becerra study.

Further analysis within the physician group by specialty revealed a consistent pattern of significant differences in mean seriousness ratings. This supports the notion that training and work role effect evaluation of seriousness. Surgeons consistently rated categories as less serious than medical students or pediatricians (significantly so in areas of minor neglect and lifestyle/values, and approaching significance for sexual abuse). The focus of surgeons is even more "biomedical" and less holistic than is true for pediatricians and medical students, and their contact with parents is apt to be minimal.

Caution must be raised in considering the implications of these professional differences. How the ratings relate to actual behavior in clinical decision-making cannot be determined from this study. However, a 3-month review of contacts by hospital professionals with the child abuse consulting team and observation of the team's weekly update meeting indicate that there may be some behavioral correlates to the response tendencies found in this study.4 Social workers are most likely to refer cases to the team for consultation. Pediatricians were the second most likely group to refer. In many cases, however, the latter group responded reluctantly, after prodding from social workers and nurses. Contacts from psychologists and psychiatrists to the team were quite infrequent. A more systematic study of hospital practices in recognition and management of possible child maltreatment cases, through review of case records and reports filed, could shed more light on the association between the attitudes reflected in this study and actual behavior.

A second caution in drawing implications about professional differences from the current study stems from examining the absolute differences between group means. While differences are significant, they are small in actual magnitude, and ratings tend to be within the same range of seriousness. Hence, there appears to be a good deal of agreement among the professional groups.

Personal Variables and Seriousness Ratings

Differences in factor ratings across professional groups by the personal variables of parenthood status, years of experience, and prior involvement with child abuse cases were not significant. Sex differences were significant on all categories except parental sexual preference, but sex was confounded with profession. Personal variables were also examined within professional groups. Women rated incidents as more serious than men, and parents gave higher ratings than did non-parents. The correlation between number of years of experience in one's profession with seriousness ratings of maltreatment categories was significant only for nurses, and the relationship was negative. Perhaps this reflects a generational difference, or a desensitization effect related to many years of inpatient experience. Thus, profession appears to be the variable that best accounts for rating differences.

CONCLUSIONS

A number of conclusions can be drawn from the findings of this study.

1. Hospital professionals discriminate among child maltreatment categories and agree on the rank ordering of their seriousness for the welfare of the child.In conjunction with Giovannoni and Becerra's (1979) findings, this suggests that definitions of child maltreatment used in clinical practice, state reporting laws, and research studies could be made more precise. Clear guidelines for practice according to maltreatment category might be developed by specifying circumstances meriting intervention.
   
2. Parental sexual preference does not belong in the domain of child maltreatment. Professionals were in agreement in assigning ratings at the "non-serious"; end of the scale to this category. Parental sexual preference was distinct from other lifestyle items, and should not be grounds for protective intervention.
   
3. Significant differences in seriousness ratings by profession occurred in all categories except severe physical abuse. These differences are best accounted for by work roles and training. Front-line nurses and social workers give the highest seriousness ratings, and psychiatrists give significantly lower ratings. Whether these differences reflect a difference in response tendency or a difference apparent from the level of professional involvement cannot be determined from this study.
   
4. Despite differences found among professionals in ratings of seriousness, there was also much consensus. The absolute differences in mean ratings were generally small. The ranking of maltreatment categories by seriousness was almost identical. Hence, the emphasis must be on consensus among hospital professionals in evaluating incidents of child maltreatment. Further study should focus on differences and similarities in behavior among professionals when recognizing and responding to cases of possible child maltreatment, as well as analyzing the relationship between attitudes and behavior. Behavioral differences could be examined indirectly through case record review or more directly through observation of actual practice. Correlations between attitudes and behavior might be examined in a variety of ways. For example, in addition to rating seriousness of incidents for the welfare of a child, professionals might be asked to indicate what action should be taken. Finally, it would be useful to systematically investigate what specific information effects the evaluation of seriousness by different disciplines (e.g., nature of the consequence, parental characteristics, child characteristics, family socioeconomic status, ethnic group). Understanding interdisciplinary differences and similarities in attitudes and behavior in these cases would facilitate better communication among professionals, and improve clinical practice.

NOTES

¹In the 1983 national statistics of reported child abuse cases (AHA, 1983), only 3.2% involved a "major physical injury," 18.5% involved "minor physical injury," and 5.2% involved an injury of unspecified severity. The most commonly reported type of maltreatment is "deprivation of necessities." The less obvious case may well be the norm.

²Physicians level of training:

³When variances between groups were significantly different, Satterthwaite's approximation was used.

⁴Boston Children's Hospital interdisciplinary child abuse team has been operating since 1970. At the time of this study it was composed of two social workers, a psychiatrist, a psychologist, a nurse, a pediatrician, and a hospital attorney. The team acts as a consulting group and becomes active only when a case is referred to them for review and advice on case management, including whether to file a report to the state protective service agency. Most referrals to the team are inpatient cases. On most hospital divisions, any of the involved professionals can refer a case to the team; on some, the referral has to come from the physician in charge of the case.

REFERENCES

1. American Humane Association (1983). Annual Report, Highlights of official child neglect and abuse reporting. Denver, Co.
2. Bourne, R., & Newberger, E. H. (1980). Interdisciplinary process group in the hospital management of child abuse and neglect. Child Abuse and Neglect, 4,137-144.
3. Daniel, J. H. (1985). Cultural and ethnic issues: The black family. In E. H. Newberger and R. Bourne, (Eds.), Unhappy families: Clinical and research perspectives on family violence (pp. 145-154). Littleton, MA: Publishing Sciences Group.
4. Elmer, E. (1960). Abused young children seen in hospitals. Social Work, 5, 98-102.
5. Erlanger, H. S. (1975). Social class differences in parents' use of physical punishment. In S. K. Steinmetz and M. A. Straus (Eds.), Violence in the family. New York: Dodd, Mead & Co.
6. Gelles. R. J. (1973). Child abuse as psychopathology: A sociological critique and reformulation. American Journal of Orthopsychiatry, 43, 611-621.
7. Gelles, R. J. (1975). Community agencies and child abuse: Labelling and gatekeeping. Paper presented at the Conference on Recent Research on the Family, sponsored by the Society for Research in Child Development, Ann Arbor, MI.
8. Giovannoni, J. M., & Becerra, R. M. (1979). Defining child abuse. New York: Free Press.
9. Hampton, R. L., & Newberger, E. H. (1985). Child abuse incidence and reporting by hospitals: Significance of severity, class, and race. American Journal of Public Health, 750 56-60.
10. Helfer, R. E. (1975). Why most physicians don't get involved in child abuse and what to do about it. Children Today, 4, 28-31.
11. Hill, D. A. (1975). Emotional reaction to child abuse within a hospital setting. In N. B. Ebeling and D. A. Hill (Eds.), Child abuse: Intervention and treatment. Acton, MA: Publishing Sciences Group.
12. Katz, M., Hampton, R. Newberger, E., Bowles, R., & Snyder, J. Returning children home: Clinical decision making in cases of child abuse and neglect. American Journal of Orthapsychiatry, 56, 253-262.
13. Kaufman, I. (1983). The professional use of the self: Transference and countertransference. In N. B. Ebeling and D. A. Hill (Eds.), Child abuse and neglect. Littleton, MA.: Publishing Sciences Group.
14. Korbin, J. (1977). Anthropological contributions to the study of child abuse. Child Abuse and Neglect, 1, 7-24.
15. "The Last? Resort" (1985). Newsletter of the Commitee to End Violence Against the Next Generation, Inc. 14.
16. Nagi, S. Z. (1977). Child maltreatment in the United States. New York: Columbia University Press.
17. Nalepka, C., O'Toole, R., & Turbett,j. P. (1981). Nurses and physicians' recognition and reporting of child abuse. Comprehensive Pediatric Nursing, 5, 33-38.
18. Newberger, E. H., Rosenfeld, A. A., Hyde, J. N., & Holter, J. C. (1978). Child abuse and child neglect. In R. A. Hoekelman, S. Blauman, P. A. Brunell, S. B. Friedman, & H. M. Seidel (Eds.), Principle, of pediatrics: Health care of the young. New York: McGraw-Hill.
19. Rosenfeld, A. A., & Newberger, E. H. (1977). Compassion vs. control: Conceptual and practical pitfalls in the broadened definition of child abuse. Journal of the American Medical Association, 237, 2086-2088.
20. Rosenzweig, H. D. (1982). Some considerations in the management of child abuse: A psychiatric consultant's perspective. In E. H. Newberger (Ed.), Child abuse. Boston: Little Brown.
21. Rosenzweig, H. D. (1985). Psychiatric issues and the physician's role in reporting child abuse. In E. H. Newberger and R. Bourne (Eds.), Unhappy families: Clinical and research perspectives on family violence. Littleton, MA: Publishing Sciences Group.
22. Straus, M., Gelles, R., & Steinmetz, S. (1980). Behind dosed doors: Violence in the American family. Garden City: Anchor/Doubleday.
23. Turbett, J. P., & O'Toole, R. (1980). Physicians' recognition of child abuse. Paper presented at the Annual Meeting or the American Sociological Association, New York, N.Y.

Acknowledgments

This research was supported by a grant from the National Institute of Mental Health (I TO 1 MH 15117 IA2 CD). A version of this paper was presented at the Massachusetts Psychological Association Meeting, Boston, June, 1982. The authors gratefully acknowledge the assistance of Wyatt Jones in study design; Terry Tivnan, Joanie Spitz, and Russell Phillips in data analysis; and the Family Violence Seminar in discussion of results.

Web design by flyte new media
email Web Master