I dedicate this thesis to my parents, Marietta and Anderson Scott, Jr., and my brother, David Todd Scott. It is with their love and support that I have navigated my educational journey.

I also dedicate this thesis to all those pioneers who came before me, and fought battles that I will never have to fight, so that I could take so much for granted.

RETURN TO INDEX

I would like to express my appreciation to the following individuals:

Dr. Lysle E. Johnston, Jr., who allowed me to be a part of the University of Michigan African American Recall Study. I feel that the UMAARS is a vitally important project, and col-laboration on this project was an experience that I will not soon forget;

Dr. James A. McNamara, Jr., my fellow University of Cali-fornia at Berkeley alumnus and Miami Triad brother. I thank him for his counsel and friendship;

Drs. Burton L. Hagler and Paul W. Reed, the other two cor-ners of the "Miami Triad of UMAARS Research." I thank them for their companionship and aid in our research travels, and I wish them the best in their careers;

Dr. Ken Guire for his expert statistical assistance in the preparation of this thesis;

Wendell W. "Trey" Campbell, III, a University of Michigan dental student who took interest in this project and assisted me in some of the painstaking labor;

Drs. Caroll-Ann Trotman, R. Angela Wandera, and Howard L. Tingling, members of my thesis committee. I thank them for their time and interest in this project;

Drs. Howard Tingling; Richard Kulbersh, University of De-troit; Alan Borislow, Albert Einstein Medical Center; Thad Champlin, University of the Pacific; Walter Buchsieb, Ohio State University; Orhan Tuncay, Temple University; Kymberly Higgins-Barber and Robert Vanarsdall, Jr., University of Pennsylvania; Mark Hans, Case Western Reserve University; Jerome Superstine; David Baru; John Lupini; and Eli Berger for allowing us to use patient records and recall patients to their respective clinics to obtain data for this study;

Finally, the sources of funding without which this study would not have been possible: National Institute of Dental Research Grant DE08716 and a University of Michigan School of Dentistry LeGro Fund Grant.  

RETURN TO INDEX

INTRODUCTION

In the 1960s, Americans of primarily African descent started calling themselves "Black" and declared that it was beautiful. Accordingly, there began a process wherein Blacks/African Americans began to look at themselves uniquely in terms not only of politics, but also esthetics. From "dashikis" to "afros," many African Americans began to feel that they wanted simply to look like themselves. The Euro-pean standard no longer was accepted without question.

Understandably, some of those new ideas on ethnic beauty would be applied to facial esthetics. If "Black is beauti-ful," certainly no one would want to be robbed of this beauty out of ignorance. As a result, and perhaps unexpectedly, the practice of orthodontics has been brought into the fray. Af-ter all, few patients walk into an orthodontist's office ex-pecting their treatment to become an issue of race. It may become just that, however, if a patient's own sense of beauty is altered by ill-considered, inappropriate orthodontic treatment.

In many fields, scientists tend to study themselves. For orthodontics, this realization means that much of what is known (or, perhaps, what is thought to be known) about facial norms and esthetics is derived from "White standards." Even today, comparatively little is known of Black cephalometric standards or esthetic preferences, not to mention the short- or long-term effects of the various treatment protocols. Practicing orthodontists must rely, therefore, on some sort of instinct to treat Black patients, or alternatively, decide to ignore the possibility of differences altogether. Clearly, this policy of ignorance or denial could well lead to undesirable consequences. More to the point of this study, it may affect patient care at one of the most basic levels of treatment planning, the decision whether or not to extract teeth.

The benefits and effects of extraction and non-extraction orthodontic treatment frequently are debated, and, from time to time, one approach or the other is in fashion. Today, non-extraction techniques are in the ascendancy, per-haps as a result of a popularly held belief that the extrac-tion of premolars to alleviate crowding or to correct bimax-illary protrusion may change the facial profile so much as to "ruin the face" by creating an unattractively concave ("dished in") profile.

The study of esthetics and preferences is an inherently subjective pursuit. This thesis, however, will pursue an ob-jective comparison of the way the faces of African-American patients are changed esthetically by extraction and non-extraction orthodontic treatment, and how that change is viewed by the patient as well as other observers. Such ob-jective knowledge is important, considering that extraction orthodontic treatment is utilized far more commonly than the non-extraction alternative for African-American patients. If extraction treatment (or the non-extraction alternative, for that matter) causes harm or displeasure, then clearly there is a need to re-think this approach to treatment.

RETURN TO INDEX

REVIEW OF THE LITERATURE

America grew by an influx people from far away places. Non-natives came first from northwestern Europe or were brought from the western coasts of Africa. Their physical differences were striking. People thus embraced as a matter of course the concept of "race" and categorical racial dif-ferences (Brace, in press). Today, it is considered a truth almost too obvious to mention that to belong to a given race makes a person physically different from a person of another race, and, by extension, that "Black" faces look different than "White" faces.

What Makes An African-American Face "Black?" Although the actual scientific definition, delineation, or separation of races may be difficult at best, it cannot be denied that there are obvious differences apparent to just about any observer. When anthropologists excavate human re-mains, they may try to identify the race of a subject by "skull racing" (Rhine, 1990). This technique of racial iden-tification incorporates observation and measurement of iden-tifiable features on the human skull. Rhine proposed a modi-fication of this technique, "non-metric skull racing," which he believed would make the racial identification of skulls more accessible. For this purpose, he listed a number of skeletal features that distinguish the skull of American Cau-casians from American Blacks.

Rhine described the Caucasian skull (Fig. 1) as gener-ally having a sloping orbit, depressed nasion, retreating zy-gomatic arches, no prognathism, a bilobate prominent chin, and an undulating mandible. By way of contrast, the Black skull tends to have rectangular orbits, a slight nasal de-pression, vertical zygomatic arches, pronounced prognathism, a blunt and vertical chin, and a straight mandibular border. What is apparent is that the distinguishing features of Blacks and Whites generally involve the middle and lower face. Richardson (1980) seconded these observations from an orthodontic perspective when he said that the "parameters of the face that are closer to the alveolar and dental areas show the greatest differences among ethnic and racial groups." It is in this area that orthodontic change is most probable and most apparent. Thus, if there is discernible difference between patient groups in this region of the face, it stands to reason such differences should be taken into consideration when establishing orthodontic treatment goals.

Cephalometric Measures in Orthodontics The study of esthetics may be inherently subjective, but the practice of orthodontic treatment planning is not. Plas-ter casts give practitioners a tangible means of examining

a

b

Figure 1. Depiction of the typical (a) African-American and (b) European-American skulls. From Rhine, 1990. and quantifying the occlusal relationships of teeth; however, it was not equally possible to examine directly the existing skeletal relationships on a living patient until the intro-duction of roentgenographic cephalometry to orthodontics by Broadbent (1931).

In 1948, Downs took a first step toward standardized cephalometric analysis by suggesting a set of ten quantita-tive measures he thought clinically meaningful to orthodon-tists. To develop normative data, he measured these ten characteristics in a group of 20 Caucasian subjects, about equally divided in gender, 12 to 17 years old, who possessed "excellent" occlusions. For each of the five skeletal and five dental measurements (Table 1), he determined the minimum and maximum values and means for his sample. Once these cephalometric values were established, they provided "norms" to which individual patients could be compared. Clearly, however, this small sample cannot be equally valid for pa-tients of other races. Three years later, Cotton and associ-ates (1951) addressed just this issue with regard to African Americans.

Racial Differences In Cephalometric Measures Cotton (in his portion of a paper in which Wendell Wylie summarized three separate studies) duplicated Downs' study as closely as possible on 20 Blacks from the San Francisco Bay Area. His sample was equally divided as to gender and ranged in age from 11-34 years. Cotton's data differed

TABLE 1 Downs Analysis: Normative Values

Measure Whites Blacks

Cotton Altemus

Facial Angle 87.9 87.25 85.7 Angle of Convexity 0.0 9.6 9.7 A-B Plane Angle -4.7 -7.7 -6.3 Mandibular Plane Angle 21.9 27.25 28.8 Y-Axis 59.3 63.3 63.4 Cant of Occlusal Plane 9.2 11.8 10.7 Inter-incisal Angle 135.4 123.0 119.2 Lower-1 to Occlusal Plane 14.5 22.5 27.3 Lower-1 to Mandibular Plane 1.5 6.6 9.8 Upper-1 to A-P Plane (mm) 3.1 8.5 10.4

* After Cotton and associates, 1951; and Altemus, 1960.

TABLE 2 Comparison of Black Means to Caucasian Norms

Mean Measure

Black White

MPA 30.6 26.4 FMIA 49.4 56.4 IMPA 100.0 97.3 Upper-1 to Lower-1 113.8 126.8 Lower-1 to NA (mm) 7.4 3.0 Upper-1 to NA 24.1 23.2 Lower-1 to NB 11.4 5.4 Lower-1 to NB (mm) 36.7 27.3 SND 75.8 76.0 SNA 84.7 81.0 SNB 79.2 78.2 ANB 5.5 2.8 GoGn-SN 38.2 32.0 Upper-1 to SN 108.9 103.8

* After Drummond, 1968.

significantly from Downs' values in many areas, particularly in terms of a higher mandibular plane angle and increased in-cisor protrusion (Table 1).

With the first cephalometric norms for African Americans now established by Cotton, others would refine his values. Altemus (1960) studied a sample of Blacks whose ages--12-16--were closer to, and more representative of, the average or-thodontic patient. His sample consisted of 80 African-American children (40 boys and 40 girls) with complete perma-nent dentitions, except for third molars.

Altemus found that this group had a skeletal pattern al-most identical to that of Cotton's group, but with a more variable dental pattern. In the end, however, he felt that the differences were relatively small, and that, overall, the two groups were "extremely close" when differences in size and makeup of the samples were considered. When Altemus com-pared his findings to the norms reported for Whites, he, too, found that there were definite differences. In terms of overall absolute head and face size, American Blacks were larger than American Whites. Also, he noted a greater prog-nathism in Blacks; but, because the facial plane was similar in both races, he felt it was a manifestation of the dental, rather than skeletal, relationships.

Another study of Black children was published eight years later by Drummond (1968). He studied 40 Black subjects at Baylor University who ranged in age from 8 to 23 years old and were thought to have both acceptable occlusions and faces that were not deformed in any way. This study used a cepha-lometric analysis that included parts of both the Alabama Analysis (Taylor and Hitchcock, 1966) and the Tweed Triangle (Table 2). Drummond was able to show significant differences in the elements of the Tweed Triangle and other measurements that demonstrated steepness in the mandibular plane and procumbancy of the incisors.

Like age, geographical location was thought to influence cephalometric norms. Alexander and Hitchcock (1978) not only limited their study to children, but also to the American South. Theoretically, a regional limitation would eliminate any environmental influences that might lead to differences in genotypic expression. Alexander and Hitchcock's subjects were 50 children from predominantly Black Catholic schools in Jefferson County, Alabama. They were from 8 to 13 years of age with Class I/normal dentitions and "acceptable" anterior dental relationships and profiles. As with Drummond's study, their cephalometric data also were compared to the norms of the Alabama Analysis (Table 3).

The observations of Alexander and Hitchcock were in line with the studies of Drummond and Altemus. Most of the skele-tal dimensions of the Black children were larger than those of the Whites. Moreover, the maxillae were, on average, more anteriorly placed, and the upper and lower incisors were more

TABLE 3 Alabama Analysis: Alexander and Hitchcock's (1978) data for Blacks compared with normative standards for Whites

Mean Measure

Black White

SNA 84.7 81.0 SNB 79.7 78.2 SNA-SNB 5.0 2.8 Facial (NP to SN) 79.2 79.0 Y axis to SN 67.2 66.1 Mand. to SN 34.4 32.0 Occl. to SN 17.8 16.4 AB to occl. 91.0 90.1 Upper-1 to occlusal 53.8 59.7 Upper-1 to NA 23.7 23.2 Upper-1 to SN 108.4 103.8 Upper-1 to AP (mm) 10.5 5.9 Lower-1 to mand. 99.3 97.3 Lower-1 to occlusal 63.9 67.2 Lower-1 to SN 46.2 50.8 Lower-1 to NB (mm) 9.7 5.4 Lower-1 to NA (mm) 6.7 6.6 Overbite 3.0 3.2 Overjet 3.4 3.5

procumbant. He also noted that neither Angle's Class II nor Class III malocclusions were as common in Blacks.

After the previous descriptions of children, Kowalski, Nasjleti, and Walker (1974) turned their attention to adults. They found significant differences between Blacks and Whites based on the descriptive measures of the Steiner (1959) analysis (Table 4). Kowalski, Nasjleti, and Walker examined the cephalometric data of 244 African-American males ranging from 20 to 60 years of age from the Veterans Hospital in Ann Arbor, Michigan. This group was compared to 381 European-American males in the same age range, also from the Veterans Administration Hospital. Kowalski, Nasjleti, and Walker found, as had others, that there were significant differences in such measures as incisor angulation (more proclined), mandibular plane angle (steeper), and increased maxillary prognathism (more anteriorly positioned). They did not, how-ever, offer their findings for adults in the context of pa-tient treatment. This would be done later by Fonseca and Klein (1978).

A few years after Kowalski, Nasjleti, and Walker studied Black men, Fonseca and Klein published a comparable study of Black women. The perspective of Fonseca and Klein was that of oral surgeons, and, thus, they looked at adult subjects as being more representative of the likely orthognathic surgery patient. In their study, Fonseca and Klein examined 40 Black women who ranged in age from 20 to 30 years old. All

TABLE 4 Steiner Norms Compared with Comparable Data for Blacks and Whites*

Mean Measure Steiner

Black White

SNA 82 86.10 82.18 SNB 80 81.22 78.94 ANB 2 4.88 3.23 GoGn/SN 32 34.31 32.89 Occl./SN 14 15.48 15.09 Upper-1/Lower-1 130 126.31 136.47 Upper-1 to NA 4 4.83 3.69 Upper-1/NA 22 32.20 32.95 Upper-1 to NB 4 9.15 4.62 Upper-1/NB 25 30.14 20.56 Lower-1/GoGn 93 94.59 88.67 Upper-6 to NA 27 26.77 25.02 Lower-6 to NB 23 18.97 19.86

* Kowalski, Nasjleti, and Walker, 1974.

subjects had a "Class I molar and/or canine relationship and clinically acceptable occlusions." A group of 20 White women (18 to 25 years old) were examined for comparison.

As with previous studies, Fonseca and Klein found (Table 5) that Black women had significantly greater bimaxillary protrusion and incisor proclination. The midfacial region of Black women was shorter and the lower facial height was greater. With regard to the soft tissue, the soft-tissue na-sal tip was less prominent in Blacks, and the upper and lower lips were more protrusive in relation to the facial plane; however, there were no real differences in the lip thickness of the two groups. This finding suggests a more direct ex-amination of the soft tissue.

Susnher (1977) conducted a comparative study solely of Black profiles using photographs taken from subjects judged to be "the most attractive-looking North American blacks pre-viously investigated by the orthodontic faculty at Howard University College of Dentistry." Sushner used Holdaway's H line (soft-tissue chin to tangent of upper lip), Ricketts' E line (soft-tissue chin to the tip of the nose), Steiner's S line (soft-tissue chin to a point where the upper lip and the base of the nose meet), and a nasion-pogonion line (soft-tissue nasion to the soft-tissue chin) to measure the Black subjects' profiles, and to compare them to White standards. He found, in agreement with previous studies, that the pro-file was fuller in African Americans. Additionally, Sushner

TABLE 5 Fonseca and Klein's (1978) Cephalometric Values for Black and White Adult Women

Mean Measures

Black White

SNA 88.2 82.5 SNB 83.9 79.6 ANB 4.3 3.0 FNA 93.1 88.0 FNB 88.7 85.1 Y-Axis 65.5 67.3 SN-GoGn 32.9 32.1 N-A/A-Po 54.6/55.5 59.5/52 MPA 27.7 26.6 INA 26.1 23.1 INA (mm) 6.6 3.4 INB 38.5 23.9 INB (mm) 9.3 4.6 I/I 112.8 131.0 I-SN 114.0 105.6 IMPA 101.2 92.5 FMIA 51.1 61.0 Nasion-soft-tissue nasion 6.7 6.8 Nasal tip-facial plane 31.2 33.5 Point A-superior labial sulcus 14.0 13.9 Labrale superius-facial plane 24.9 17.8 Labrale superius-labial edge of maxillary incisors 11.2 11.1 Labrale inferius-labial edge of mandibular incisor 13.1 12.8 Labrale inferius-facial plane 23.2 15.8 Point B-inferior labial sulcus 13.2 11.9 Pogonion-Point C 14.0 12.1

suggested his own set of African-American standards based on the profiles he studied (Table 6).

It can be seen from the studies presented here that a protrusive dentition and skeletal relationship are hallmarks of the African-American face. Given this fact, it is under-standable that these characteristics also may affect the way each group looks upon facial esthetics. The Influence of Race on the Appraisal of Beauty In one of the most commonly referred to studies of fa-cial esthetics, Peck and Peck (1970) began with a discussion of the earliest renderings of the Egyptians (and how they re-lated to the citizenry of the era and area--an African and European mixture), the combination of esthetic and spiritual philosophies by the Greeks and Romans, and then tied those into the practical, everyday concerns of the clinical ortho-dontist:

The philosophers felt that beautiful creations respected certain geometrical laws, since true beauty necessarily displayed harmony. As harmony was the "due observance of proportions," it seemed reasonable to assume that these proportions were fixed quantities. Of course, "aesthetics" as introduced by the Greeks and expounded by modern philosophers encompasses more than simply physical or natural beauty. The beauties of human emo-tion, existence, and experience are manifestly important to the aesthetician. (At this point the orthodontist must recoil. He is at a lost to program these tenuous "gray areas" into the scientific language. So the or-thodontist rightfully has streamlined "aesthetics" into "esthetics" and has limited its scope to only those cri-teria tractable to objective analysis.)

TABLE 6 Mean Values for Sushner's (1977) Cephalometric Soft- Tissue Standards for African Americans (in mm)

Measurement Male Female

Upper lip to Ricketts E plane 0.33 0.47 Lower lip to Ricketts E plane 2.00 1.10 Upper lip to Steiner S plane 5.50 4.30 Lower lip to Steiner S plane 5.00 3.90 Subnasale to Holdaway H line 7.80 6.20

In one form or another, it is this need to put the in-herently subjective evaluation of beauty into objective, sci-entific terms that is seen in virtually all studies of ortho-dontic esthetic appraisals. This process is more difficult if one has to consider the influences of race on this percep-tion, and how it might cause people (orthodontists and their patients, for instance) to look at beauty differently.

Martin (1964) examined the differences in the perception of beauty between Blacks and Whites by showing 10 selected pictures of Black women to 50 White and 50 Black college-age American men, and another 50 Black Nigerian men of similar age. The observers were asked to rank the photos in terms of attractiveness. Martin found that there was consistent agreement between American Blacks and Whites (the Nigerians disagreed with both groups, it should be noted) as to which faces were the most attractive. Further, the two American groups ranked highest the faces deemed the most Caucasian in appearance by another group of judges. From that, Martin concluded that a "Caucasian facial model" was the "single cultural model in polyracial American society." This strong assertion suggests that European-Americans and African Ameri-cans perceive Caucasian faces to be more attractive than Black faces.

Indeed, even part of Peck and Peck's work would seem to confirm this assertion. Their study compared the cephalomet-ric measurements of people already considered beautiful by society at large (movie stars, beauty queens, etc.) to com-monly-accepted norms. They found neither of their two "beautiful" Black female subjects seemed to "exhibit the an-thropological characteristics of her race." Each seemed to "posses many Caucasian-type features." Thus, Martin seems to have been correct in suggesting that all Americans look at beauty through a White standard; however, the other half of the Peck and Peck study appears to conflict with this conten-tion.

With regard to the "beautiful people" who were White, Peck and Peck found that their profiles were significantly fuller than White cephalometric norms. Therefore, it is not completely accurate to say (as Martin did and Peck and Peck reiterated) that there is universal preference for a "Caucasian facial model." What is shown most convincingly is that the most attractive faces may not be as flat or as full as their respective norms. This observation is something to be taken into consideration when orthodontic treatment is un-dertaken to change these characteristics. The fact that a person is Black or White may not mean that he or she finds most attractive African-American faces as full as the Black norm or European-American faces as flat as the White norm. Influence of Orthodontic Training on the Appraisal of Esthetics As with race, orthodontic training has been considered as an influence in the way one may judge facial esthetics. The results, implied and explicit, have been mixed, but the impression that orthodontists generally agree with layper-sons, seems to be the prevailing assessment.

To some (Spadafore, 1995; Oynick, 1988), it may seem that Peck and Peck (1970) showed that orthodontists are in disagreement with the general public as to what kind of faces are most esthetically pleasing. This interpretation, how-ever, does not come from a direct comparison between the ap-praisals of orthodontists and laypersons, but rather from Peck and Peck's observation that cephalometric measures of media stars were different (fuller) than customary cepha-lometric norms.

A more direct comparison comes from Prahl-Andersen and co-workers (1979), who asked parents, orthodontists, and gen-eral dentists to rate line drawings of profiles and intraoral photographs. The groups were to rate each profile and photo-graph as being normal; deviating from normal, but not needing orthodontic treatment; or deviating from normal and needing orthodontic treatment. The results indicated that the ortho-dontists and general dentists generally agreed with each other, but not with the parents. The orthodontists and gen-eral dentists were more likely than the parents to think that the subject needed orthodontic treatment. This study indi-cates that there are perceived differences between profes-sionals and the public.

More often, studies have reflected an observation that orthodontists and laypersons generally have the same idea about what is attractive. Before Peck and Peck (1970), Riedel (1957) studied beauty contest winners (Seattle Seafair princesses) and determined that those judged to be most at-tractive were, indeed, similar to established cephalometric norms. By the same logic previously mentioned for Peck and Peck, this would imply that orthodontists and laypersons are in agreement. Other studies are more direct in their com-parisons between orthodontists and the public.

In a study on "facial harmony" from the Netherlands, Cox and van der Linden (1971) asked 10 orthodontists and 10 laypersons to evaluate the silhouette photographs from 241 females and 186 males. The results showed no significant difference between the orthodontists and laypersons. Each group had similar ideas as to which profiles were the most attractive.

Cox and van der Linden's conclusions would be supported by the findings of De Smit and Dermaut (1984) in Belgium. De Smit and Dermaut used silhouette profiles to compare the re-sponses of those who had orthodontic training and those who did not. When they did, they found no significant differ-ences in the way the profiles were evaluated. The groups of orthodontists and non-orthodontists similarly ranked the given selection.

Orthodontists and laypersons also have been shown to share similar preferences when the profiles of patients are changed in some way. Dunlevy and associates (1978) showed pre- and post-treatment photos of orthognathic surgery pa-tients to panels of laypersons, orthodontists, and oral sur-geons. When the panels were asked to rank the subjects by the amount of treatment improvement, there was general agree-ment among the groups. Each group indicate the same patients (those patients with the most surgical advancement) as being most improved.

Farrow, Zarrinnia, and Azizi (1993), on the other hand, did not surgically treat patients, but asked groups of Blacks, Whites, orthodontists, and general dentists to evalu-ate digitally altered photographs. They found that all groups responded most favorably to the same type of profile. Professional training did not lead to differences in overall preferences. Judges from different educational backgrounds given the choice between, say, "good" profiles and "bad" pro-files, will most likely choose the same way. What may sepa-rate judges, however, is the strength of their preferences.

Studies by Oynick (1988) and Spadafore (1995) were simi-lar in that they asked groups of laypersons and orthodontists to evaluate pre- and post-treatment profiles. In Oynick's study, it was found that there were no significant differ-ences in the perceptions of esthetic improvement by orthodon-tists and laypersons. Each group judged the same profiles to be most pleasing--those profiles of patients who had had a significant reduction in lip procumbancy. Spadafore, too, found similarities in the absolute choices made by orthodon-tists and laypersons (all liked treated faces better than non-treated faces, and faces treated by non-extraction better than faces treated by extraction); however, there were dif-ferences in the magnitude of the judges' responses. The scores of the orthodontists showed the greatest variance be-tween non-extraction and extraction profiles (the scores be-come markedly lower for extraction): they rated the extrac-tion profiles lower than any other group (oral surgeons, White laypersons, and general dentists), save for Black laypersons. Thus, Spadafore showed that there are general similarities between those with orthodontic training and laypersons. He also showed that interaction between educa-tion and extraction can influence one's response. Thus, the effect that premolar extraction has on esthetics must be con-sidered as well as the influence of orthodontic training. Influence of Orthodontic Treatment on Profiles

Effect of Extraction
Premolar extraction treatment is the basic strategy most commonly applied to Blacks. The effects of this strategy are, therefore, of great importance to a study of facial es-thetics in African Americans because it affects directly those characteristics that give a face its "Black" character-istics (i.e., bimaxillary protrusion and steep mandibular plane inclination).

The extraction of premolars (usually the four first pre-molars) permits the retraction of upper and lower incisors. Further, incisor retraction will be followed by a reduction in the fullness of the lips, and a subsequent straightening of the profile. Just how much reduction in lip protrusion will follow incisor retraction (i.e., the ratio of incisor retraction to lip reduction), or whether significant incisor retraction could lead to an esthetically undesirable flatten-ing of the face, still is the subject of debate.

Consider Hershey's (1972) study of incisor and lip re-traction. He gathered a sample of 36 mature Caucasian fe-males whose average age at the beginning of treatment was 20.3 years. He measured the movement of four hard-tissue points, Point A, upper incisor point (the most anterior point of the crown of the upper incisor); Point B lower incisor point (the most anterior point of the crown of the lower in-cisor); and four soft-tissue points, superior labial sulcus, labrale superius, labrale inferius, and inferior labial sul-cus. He found that incisor retraction tended to produce lip retraction, but that large amounts of incisor retraction did not necessarily mean large amounts of lip retraction. Thus, the consensus that incisor retraction reduces lip prominence was sustained; however, an incisor-to-lip-retraction ratio could not be characterized.

A subsequent investigation by Roos (1977) studied a Class II, Division 1 sample of 10 boys and 20 girls, presuma-bly all Caucasian, ranging in age from 8 years to 16 years. His goal was to assess the correlation between hard- and soft-tissue changes during treatment. He found, in the face of a great deal of individual variation, an average ratio of 2.5:1 for the upper incisor to upper lip and 1:1 for the lower incisor to lower lip. This finding is in general agreement with reports from Rudee (1964) and Wisth (1974), who reported the same ratio for the lower incisor to the lower lip, and ratios of 2.9:1 and 2.5:1, respectively, for the upper lip.

In a study of Caucasian extraction and non-extraction patients (determined to be "borderline" by discriminant analysis), Paquette, Beattie, and Johnston (1992) demon-strated an upper incisor to lower lip retraction ratio of 1:1 in extraction patients. This ratio comes from linear cepha-lometric measures that show that, on average, the upper inci-sor was retracted 3.2 mm (upper-1 to NA) and the lower lip (lip to E plane) retracted 3.1 mm. Given that it was a com-parative study, they also were able to demonstrate that ex-traction treatment created a profile that differed about 2 mm in fullness from non-extraction profiles. Therefore, the re-duction in lip procumbancy due to incisor retraction was shown to be consistent and significantly greater in extrac-tion than in non-extraction treatment. A similar 1:1 ratio was found by Luppanapornlarp and Johnston (1993) in a group of clear-cut extraction patients. They also found a similar 2 mm difference in the lip procumbancy at the end of treat-ment when the clear-cut extraction patients were compared to clear-cut non-extraction patients.

One balance, great consistency can be seen in the ratio of upper incisor to lower lip retraction, and more varied re-sults with the upper incisors and upper lip. It remains to be seen, however, whether these findings can be applied equally to African Americans.

Incisor and Lip Retraction in African-American Patients
The specific investigation of incisor and lip retraction in African Americans may have begun with Garner (1974), who studied the records of a sample of 15 consecutive Black pa-tients, 10 girls and 5 boys. They ranged in age from 11 years to 14 years at the beginning of treatments averaging 24 months in length. When pre- and post-treatment records were compared, Garner observed that "Negro boys and girls . . . exhibit changes in lip posture as a result of tooth movements . . . [but] the extent of the lip change is not always pre-dictable." He also stated that the changes in lip posture did not seem to be the same in African-American boys as girls; however, his sample of boys was too small to support a more precise characterization. Having only five male sub-jects magnified the effect of an individual variation that included one patient whose lips actually became more protru-sive after incisor retraction. In the end, Garner had to fall back on the suggestion that one use the same rule of thumb for Black patients as for White: a 2:1 ratio for max-illary lips and incisors, and a 1:1 ratio for mandibular lips and incisors.

Historically, others have had an equally difficult time in finding ratios that apply directly to African Americans. Park, Kudlick, and Abrahamian (1989) examined the records of 15 African-American premolar extraction patients from Howard University. At the start of treatment, the girls had a mean age of 10 years; the boys, 13 years. Lip changes were ana-lyzed in the horizontal and vertical dimensions. Change in the nasolabial and mentolabial angles were analyzed, as well.

Park, Kudlick, and Abrahamian's sample showed an average maxillary incisor retraction of 3.87 mm; mandibular incisor retraction was considered to be insignificant because of the masking effect of mandibular growth. They found a direct re-lation between upper incisor retraction and upper lip depth (i.e., lip protrusion). Park, Kudlick, and Abrahamian did not, however, propose a clinically applicable ratio for the relationship between incisor retraction and lip protrusion, because they felt there was not sufficient evidence.

Diels and associates (1995) studied the soft-tissue changes in 30 male and 30 female African-American patients exhibiting bialveolar protrusion. In males, they found an upper lip procumbancy reduction of 1.5 mm; in females, 1.7 mm. The lower lip retracted 2.7 mm in males and 2.5 mm in females. Accordingly, they found a ratio of lower incisor retraction to lower lip reduction of 1.2:1 for males and 1.6:1 for females. The upper-incisor-to-upper-lip ratio was 3.3:1 in females; however, the ratio in males was not consid-ered relevant because the upper lip was moving forward, pre-sumably due to growth, even though the upper incisors were moving backwards due to treatment.

Many of the papers presented in this review of soft tis-sue studies, although well-intentioned, are handicapped by samples that do not support meaningful comparison, or which have inadequate statistical power. Some recent studies, how-ever, have begun to address these deficiencies.

The University of Michigan African-American Recall Study
In an attempt to address problems of sample size and un-biased comparison, a long-term recall of African-American ex-patients has been undertaken by the University of Michigan. A mixed sample of Class I and II African-American ex-patients who were determined mathematically to have been similarly susceptible to extraction and non-extraction orthodontic treatment were recalled an average of eight years post-treatment. The similarity of these patient groups allows for a relatively unbiased comparison of the effects of extraction and non-extraction fixed appliance treatment. Thus far, the data from the University of Michigan African-American Recall Study (UMAARS) have been used to explore the morphological, functional, and periodontal impact of these two treatment strategies. Indeed, it is from the cephalometric radiographs traced by Hagler (1995) that the profile silhouettes for the present study were obtained (vide infra). More to the point of this review, Hagler's characterization of profile changes are pertinent to a discussion of incisor and lip retraction in African-American patients.

In his analysis of the UMAARS data, Hagler found a high correlation between incisor retraction and the reduction of lip protrusion. On average, the upper incisors in the ex-traction group were retracted 4.2 mm; the lower incisors, 3.2 mm. This finding stands in contrast to the non-extraction group, whose upper were incisors protracted 0.5 mm and whose lower incisors were protracted 0.3 mm.

The upper-incisor retraction was followed by upper-lip retraction at a 2:1 ratio after an initial upper incisor re-traction of 1-2 mm to counteract the effects of growth and/or lip strain. Lower lip retraction, on the other hand, began almost immediately with the initial retraction of the upper and lower incisors, and occurred at nearly a 1:1 ratio with the lower incisors.

Hagler also added that, compared to Class II, division 1 Caucasian patients (Paquette, Beattie, and Johnston, 1992), the UMAARS data (largely Class I) tended to exhibit less lip movement for a given amount of retraction. Although the lower incisor retraction was similar for African Americans and Caucasians (approximately 4.0 mm for the extraction groups and 0.6 for the non-extraction groups), lower lip re-traction in the Caucasian patients was 2.0 mm greater in the extraction group and 2.5 mm greater in the non-extraction group than for African Americans. Most likely this was do a different lip musculature in Whites, which leads to greater and more immediate lip retraction when incisors are re-tracted.

It is of interest to note in the UMAARS sample that the extraction post-treatment profile differed from the non-extraction post-treatment profile by 2 mm in upper lip posi-tion and 3.8 mm in lower lip position. Hagler also wrote that the profiles of each group showed continued flattening after an average of five years of retention.

Orthodontists, therefore, can produce incisor and lip retraction, which changes, in turn, determine how profile es-thetics are changed by treatment. This control, however, must be appropriate to the need of each patient group.

Esthetic Evaluation Of Profile Changes In African-American Patients
There are relatively few studies that have investigated the esthetic impact of orthodontic treatment in African Americans, and fewer yet that deal specifically with the es-thetic consequences. In 1978, DeLoach was one of the first to address directly "the desires and opinions of North Ameri-can Blacks regarding their facial profiles." In his study, he sought to determine whether or not Black esthetic prefer-ences were consistent throughout various age groups; whether or not people had an accurate perception of their own pro-files; whether or not many would change their profiles and, if so, what modifications they would favor.

DeLoach analyzed the preferences of African Americans by way of questionnaires filled out by 224 Black females. He found a considerable uniformity in their esthetic prefer-ences, regardless of age or socio-economic background. From a selection of profiles representing different facial types, the great majority chose the same profile, one that was "straighter" than the Black norm. Further, he evaluated the faces of the subjects and found that they did not, when asked, do well at recognizing there own profile types. Also, most did not want a change in their profile appearance, but those who did cited mainly soft-tissue changes, such as a re-duction of the thickness of the lips or nose.

These findings seemed to corroborate Martin's (1964) contention that American Blacks and Whites share a common es-thetic preference with regard to female facial beauty. De-Loach, however, did not go so far as to support Martin's other contention that Blacks and Whites share a common affin-ity for a Caucasian facial model, perhaps because the face that his subjects most favored, although flatter than Black norms, was also fuller than White norms.

Thomas (1979) used DeLoach's profiles to evaluate the responses of White and Black members of the orthodontic com-munity. In doing so, he observed that Black and White ortho-dontists found the same profile most pleasing. It was a relatively straight profile with only a slight protrusiveness to the lips. Like DeLoach, he felt that this outcome was yet another an example of Martin's original proposition that Blacks preferred a straighter-than-normal profile. Unlike DeLoach, however, Thomas directly supported Martin's theory that Blacks are attracted to a Caucasian facial model.

Recently, in a departure from the use of silhouettes, Farrow, Zarrinnia, and Azizi (1993) used digitally manipu-lated photographs of African-American subjects to offer judges a photo-realistic choice of profiles. In these pic-tures, the lips had been modified in three millimeter incre-ments to give each face four different appearances. The first profile represented a Black face with White cephalomet-ric norms and lip position. The second profile was a compro-mise between Black and White norms, and the third followed Black facial norms. The fourth profile was more protrusive than the Black norm. The final tabulation showed that the profile representing a compromise between African-American and European-American norms was the most frequently selected by all groups. Subsequent data agree with this conclusion.

Polk and associates (1995) found that male and female African Americans are in general agreement as to which pro-files they prefer. The only difference was that a slightly fuller profile was preferred for Black men than Black women; however, all subject groups agreed on that as well. Also, a substantial number of respondents failed to identify cor-rectly their own profile type. Further, about one-third of the Black respondents indicated that they would like to change something about their profiles, and these changes usu-ally involved the lips, chin, or nose. Of course, this find-ing also means that two-thirds of the group liked their faces as they were. Overall, Polk and associates' work validates the argument that African Americans prefer profiles flatter than Black norms, but fuller than is typical for White pro-files. Given that profile preferences are reasonably well known, this information must then be put into the context of orthodontic treatment and the comparative evaluation of vari-ous therapeutic strategies.

Comparative Clinical Research
Comparative research has been sub-divided by Feinstein (1985) into research that evaluates cause-effect relation-ships ("impact" studies) or the quality of the process ("therapeutic trials"). The former type of comparative re-search, the impact study, may be most applicable to a study (such as the present study) that investigates the effects or changes produced by a agent or modality of treatment. Kramer (1988) further sub-divided impact studies according to de-sign; "experimental studies" (also called "clinical trials"), if they are based on performing new treatment on patients; "surveys" ("observational studies"), if they are based on treatment assignment. The present study is observational in that it does not subject patients to new clinical treatment, but rather compares retrospectively the esthetic impact that two different types of treatment have made on a sample of pa-tients selected for comparison.

Given the appropriate architecture, retrospective analy-sis can alleviate many problems inherent in the prospective clinical trial (Johnston et al., 1991; Paquette, Beattie, and Johnston, 1992; Johnston, 1995). Both the need to wait for the completion of treatment and the expenses associated with a protracted study are reduced, because all needed informa-tion can be assembled in a comparatively short time. Retro-spective analysis, however, requires considerable planning if one is to minimize the effects of bias, most particularly, susceptibility bias.

Problems with Bias
Susceptibility bias arises from the fact that subjects treated one way tend to be consistently different from sub-jects treated another (Johnston et al., 1991). Such can be the case with extraction patients who, before treatment, com-monly present with more bimaxillary protrusion and crowding than non-extraction patients. Therefore, if the extraction and non-extraction patients were different before treatment, then it is difficult to interpret differences that may exist after treatment. To minimize susceptibility bias, the UMAARS samples were defined with respect to the characteristics that prompted the original treatment assignment. To this end, the initial records (cephalograms, study models, and demographic data) were subjected to discriminant analysis, a multivariate procedure for identifying characteristics that distinguish subjects in one group from another. The characteristics, once identified, can be used to classify subsequent subjects (Kachigan, 1985).

In a retrospective comparison of the outcomes of two different types of treatment, discriminant analysis can be used to identify characteristics that might cause a subject to be assigned one treatment or the other. These character-istics then can be used to generate a discriminant function whose scores separate the subjects into two groups: "clear-cut" patients (those patients at the extremes of the curve who appear to be susceptible to only one of the two treatment alternatives), and, where the curves overlap somewhere in the center (as they commonly do), a second group similar enough to be considered equally susceptible to either treatment (Johnston et al., 1991).

It is the "equally-susceptible" patient found in the re-gion of overlap between extraction and non-extraction that comprises the UMAARS sample. Presumably, the resulting be-tween-groups similarity before treatment will allow for a relatively bias-free comparison of the effects of extraction and non-extraction orthodontic treatment--effects that are of great significance to the treatment of African-American pa-tients.

Summary and Statement of Thesis
Almost any person can recognize the phenotypic differ-ences that associate a person with one racial group or an-other. The faces of African Americans differ from European-Americans in that the lips are more procumbant and the pro-files more convex; however, this normally protrusive profile can be flattened by standard orthodontic treatment that in-volves the extraction of four first premolars, a type of treatment that features a retraction of incisors, and, most likely, a reduction in lip procumbancy. There is concern, however, that the flattening of a convex profile, although seen as an improvement in Whites, may well be perceived as less beneficial in a group for whom convexity is a perfectly normal trait. The purpose of this study, therefore, is to employ data from the UMAARS to compare the perceived esthetic impact on African-American faces of extraction and non-extraction orthodontic treatment and to examine how this per-ception is affected by the race and education of the ob-server.  

RETURN TO INDEX

SUBJECTS AND METHODS

The purpose of this study is to compare the esthetic im-pact of extraction and non-extraction orthodontic treatment on the University of Michigan African-American Recall Study sample. Esthetic impact was judged, both by the patients, themselves (who did not know they were looking at their own profiles), and by panels of African-American and European-American orthodontists and laypersons. This study is de-signed, therefore, not only to compare the esthetic impact of two types of treatment, but also to test whether the percep-tion of the esthetic impact of the treatments is influenced by race and orthodontic training. Patient Selection

There are many instances in which extraction or non-extraction is an obvious choice. This study, however, exam-ined similarly susceptible patients who do not easily fall into either choice and who thus could be justifiably treated either way. It extends on the work of Lupini (1994) and Ha-gler (1995), who identified and defined the UMAARS sample, a group determined by stepwise discriminant analysis to be similarly susceptible to extraction and non-extraction ortho-dontic treatment.

The process of patient selection began by surveying treatment records from a number of private practices and graduate orthodontic programs throughout the country. Poten-tial subjects met the following criteria: 1) African Ameri-can, based on an initial evaluation by the orthodontist and a later confirmation by the patient on a demographic question-naire; 2) availability of complete diagnostic records, in-cluding pre- and post-treatment lateral cephalograms and study models; 3) an initial Class I or Class II, division 1 malocclusion; 4) no congenitally missing teeth anterior to the third molars; and 5) treatment with an edgewise appli-ance, with or without extraction of the first premolars, com-pleted prior to 1990, thereby ensuring a post-treatment in-terval of at least five years. The selection of patients treated by a number of practitioners at a variety of loca-tions should minimize proficiency bias and enhance the "generalizability" of the results.

Initially, nearly 10,000 records were examined. Of those, 463 former patients were identified as potential sub-jects. At least three attempts were made to contact these former patients by telephone to determine their availability and willingness to present for a follow-up clinical examina-tion. They were told that they would be paid $100 for their participation if, based on an analysis of their pretreatment records, they were designated as "borderline" subjects. A total of 125 patients (85 extraction and 40 non-extraction) who met the inclusion criteria agreed to participate, if asked, and were evaluated by stepwise discriminant analysis (Fig. 2a). Based on the resulting discriminant scores, the number then was reduced by discriminant analysis to 30 ex-traction (average age, 15 years, 1 month) and 30 non-extraction (average age, 16 years, 0 months) ex-patients with "overlapping" discriminant scores (Fig. 2b). These patients would become the source of profiles used in constructing the test instruments employed here.

A major advantage of using the UMAARS sample is that initial facial form and the way each patient was physically changed by orthodontic treatment is documented fully. Moreo-ver, because of the apparent overall similarity to Richard-son's (1991) norms for 15 and 16 year olds, it is apparent that this sample is fairly representative of African Ameri-cans (Hagler, 1995). Thus, an analysis of these ex-patients should generate data that are applicable to modern African-American patients. Hagler's pre- and post-treatment tracings of the extraction and non-extraction patients show the treat-ment effects the various observers were asked to evaluate in the present study (Figs. 3 and 4).

a

b

Figure 2. Distribution of discriminant scores. a. Par-ent samples, b. Sub-samples. Upper arrows denote the mean discriminant scores for each group (i.e., group centroid). The left upper arrow is for the extraction group; the right, for the non-extraction group. The lower arrow denotes the mean discriminant score for the two groups, combined. (After Hagler, 1995.)

a.

b.

Figure 3. Averaged pre-treatment cephalometric tracings of the borderline sub-samples. Extraction, red (average age, 15 years, 1 month); non-extraction, blue (average age, 16 years, 0 months). a) Raw data, b) after a uniform 2.5% enlargement of the extraction group, which contained a higher proportion of females. Thus, the extraction group was enlarged here to facilitate visual comparison of initial facial form. The enlargement was not used in the analysis of the data. (After Hagler, 1995.)

a.

b.

Figure 4. Averaged post-treatment cephalometric tracings of the borderline sub-samples. Extraction, red (average age, 18 years, 3 months); non-extraction, blue (average age, 18 years, 8 months). a) Raw data, b) after a uniform 2.5% enlargement of the extraction group. The extraction subjects, on average, were physically group because of a higher proportion of females. Thus, the enlargement is to facilitate visual comparison. The enlargement was not used in the analysis of the data. (After Hagler, 1995.)

As may be seen, the overall profile difference in the two groups after treatment is only a few millimeters. On average, the two treatments differed by about 2 mm in upper lip position and 4 mm in lower lip position. These differences were due to lip retraction in the extraction group and incisor proclination in the non-extraction group.

Panels

In addition to the patients, themselves, four additional panels of observers were asked to evaluate the esthetic impact of extraction and non-extraction treatment on all 60 ex-patients of the UMAARS sample. These panels were comprised of African-American orthodontists (N = 67), European-American orthodontists (N = 27), African-American laypersons (N = 32), and European-American laypersons (N = 30). As with the ex-patients, ethnicity was determined by the observers' responses to a demographic questionnaire. Those who identified their race or ethnicity as other than African American or European-American, by whatever terms they chose, were excluded from the study.

The African-American orthodontists (60 practitioners and 7 residents) were recruited from the University of Michigan's Graduate Orthodontic Program, attendees at the Graduate Orthodontic Residents Program (GORP, a yearly conference of graduate orthodontic residents from throughout the country), and from the practitioners listed in the Directory of Orthodontists of African Descent (Russell, 1994). This directory is a comprehensive listing of Black orthodontists and orthodontic residents in the United States and other countries. Thus, every Black resident and orthodontist practicing within the United States was contacted by mail and asked to participate. Approximately 27% of those who were contacted responded, and all who responded were included in the sample. Only in the group of African-American orthodontists did some who received booklets choose not to participate. The panel of European-American orthodontists (16 practitioners and 11 residents) came from the faculty who teach in the graduate clinic of the University of Michigan Graduate Orthodontic Program and resident participants in GORP.

Many of the Black and White laypersons were recruited from Sacramento (California) City College. They were members of Professor Marietta Scott's English Composition class, which was about 10% African-American and 85% European-American at the time. Every member of the class was asked for voluntary participation. Other Black laypersons came from the McNichols Dental Clinic (Detroit, MI). At this dental clinic, parents and patients in the waiting room were asked to participate. Other White laypersons were recruited from the staff of the University of Michigan Graduate Orthodontic Program, all of whom were asked to participate. Additionally, available laypersons with whom the present re-searcher came into contact (both Black and White) were asked to participate. All who were willing to participate were included. None of the panel members were compensated for their participation. Table 7 summarizes the panelists.

Profile Evaluation

Test Instruments

Profiles

Silhouettes of the 60 ex-patients were used as test objects to compare the esthetic impact of extraction and non-extraction orthodontic treatment. The profiles of Hagler's pre- and post-treatment tracings (executed with 2H lead on 0.003" frosted acetate, Ortho/trac, RMO, Inc., Denver), were digitized according to a customized 25 point digitization regimen (Fig. 5, adapted from the standard 71-point regimen, Dentofacial Planner, version 5.23, Dentofacial software, Toronto, Ontario, Canada). After digitization, they were plotted on white paper and used in the construction of a series of profile sheets, one per patient. For each profile sheet the pre- and post-treatment pro-files were placed side by side in random order as decided by a coin toss. If necessary, the profiles were adjusted in size (with the use of a photocopying machine) so that the effects of growth did not make one appear larger than the other. The various observers (ex-patients, panelists) were then asked to choose which profile they preferred (vide infra).

Table 7 Patient Sources

Panel Description

BO BL WO WL

Russell Directory 58

Michigan Grad Program 1 16 5

GORP* 7 11

McNichols 18

Sacramento City College 11 19

Other† 1 3 6

Total 67 32 27 30

Male 51 13 19 11

Female 16 19 8 19

* This group includes orthodontic residents who may come from the University of Michigan Graduate orthodontic Program and/or are listed in the Russell Directory

† This category denotes individuals not belonging to any particular group, but approached by the researcher for participation.

Figure 5. Profile digitization regimen. Adapted from the standard 71-point regimen, Dentofacial Planner, version 5.23, Dentofacial Software, Toronto, Canada. Point 1, Gla-bella; 2, a landmark halfway between Glabella and Soft Tissue Nasion; 3, Soft Tissue Nasion; 4, a landmark at the junction of the inferior limit of the concavity of Soft Tissue Nasion and the dorsum of the nose; 5, Nasal Dorsum; 6, a landmark at the junction of the dorsum and tip of the nose; 7, Superior Nasal Tip; 8, Pronasale; 9, Inferior Nasal Tip; 10, Columella; 11, Subnasale; 12, Superior Labial Sulcus; 13, Labrale Superius; 14, a landmark between Labrale Superius and Stomion Superius; 15, Stomion Superius; 16, Stomion Inferius; 17, a landmark between Stomion Inferius and Labrale Inferius; 18, Labrale Inferius; 19, a landmark between Labrale Inferius and Labiomental Fold; 20, Labiomental Fold; 21, a landmark between Labiomental Fold and Soft Tissue Pogonion; 22, Soft Tissue Pogonion; 23, Soft Tissue Gnathion; 24, Soft Tissue Menton; and 25, Cervical Point.

Visual Analog Scale
To measure the strength of each observer's profile pref-erence, a Visual Analog Scale (VAS) was placed at the bottom of each profile sheet. The VAS, as validated by Price and associates (1983), is a tool that gives a measurable, numeric score to a largely subjective assessment. The scale does not have numbers or a space for a written evaluation. Instead, it uses a graphic representation of approval or disapproval, which is later given a numerical value. Such a graphic rep-resentation may take the form of happy and unhappy faces, partially full water glasses, or a ladder to demonstrate various levels of pain or approval. In the present study, the VAS was structured as a 100 mm line anchored by "SAME" on the left side and "VERY MUCH BETTER" on the right.

Once an observer had chosen the profile he or she pre-ferred, he/she was asked to rate the strength of this prefer-ence by placing a mark on the VAS line; closer to "SAME" if there was no strong preference, or closer to "VERY MUCH BET-TER" if the preference was strong. The position of the mark then was measured in millimeters from the left anchor. If the participant chose the post-treatment profile, the mil-limetric score was given a positive sign; if the participant chose the pretreatment profile, the score was given a nega-tive sign. Thus, the impact of treatment was, in effect, measured along a 200 mm scale from -100 to +100.

The measurements were adjusted to compensate for the process of photocopying, which shrank each VAS line about 5%. All lines were measured as printed, and the actual shrinkage was used to adjust each score:

VAS score x (100 / VAS length) = actual VAS score.

Subject Evaluation
After signing a Consent for Experimental Procedure form (Appendix II), each of the 60 ex-patients was asked to evalu-ate only one set of profiles--his or her own. They were not told that the profiles they were judging were their own, and no patient indicated that he or she recognized the identity of the profiles. All ex-patients completed their sheet in the company of the present investigator. Verbal instructions were given on how to use the VAS, and any questions that arose were answered at once. Panel Evaluations In addition to the self-evaluations by the patients, four panels of observers were asked to evaluate all 60 pa-tient profiles assembled into a booklet of profile sheets (Appendix I), as well as to fill out a demographic question-naire seeking basic information about race, education, and gender. Personal supervision was not usually possible when the panels evaluated the booklets, most of which were dis-tributed by mail. The materials posted to each panelist thus included written instructions and a telephone number they could call (collect) if further instruction was needed. Only one panelist utilized this aid.

Despite the written instructions, some panelists did not mark the VAS properly. Indeed, six booklets were so incom-plete that they had to be excluded altogether; however, be-cause of the danger of distorting the data by introducing ex-clusion bias, every effort was made to include all interpret-able responses. Thus, if a judge skipped over only a few profiles (most likely the result of turning over more than one page at a time), the remainder of their assessments were included in the data. If a panelist marked the VAS with an "X," instead of a hash mark, then the measurement was taken from a perpendicular erected through the spot where the arms of the "X" crossed each other.

Statistical Analyses
Means and standard deviations for VAS scores were calcu-lated with a Macintosh-based spreadsheet (Microsoft Excel, version 5.0, Microsoft Corporation, Seattle, WA). Graphs that illustrate the patient distribution with respect to the discriminant analysis results, ex-patient and panel scores, and a scattergrams comparing panel data also were done with the Macintosh-based spreadsheet.

The mean scores of the groups of extraction and non-extraction ex-patients were compared to each other by way of completely-randomized t-tests. A repeated-measures ANOVA was used to analyze the panel data. Both the repeated-measures ANOVA and the t-tests were performed with a PC-based statis-tical software package (BMDP2V, BMDP Statistical Software, Inc., Los Angeles, 1988).

The data from each panelist were averaged to produce a single score for the extraction treatment and a single score for the non-extraction treatment. This "rater-based" analy-sis preserved the ratings of each observer, and, at the same time, permitted an examination of the effects of race and or-thodontic training as they affect the evaluation of extrac-tion and non-extraction profiles. Thus, the data analysis took the structure shown in Table 8. Repeated-measures ANOVA was used to identify significant differences between groups: Blacks and Whites, orthodontists and laypersons, extraction and non-extraction. In the event of interaction, simple main effects were analyzed further by way of 1-degree-of-freedom ANOVA.

For each of the 60 patients, the correlation between av-erage scores from pairs of panels was calculated as a measure of between-panel covariance. Correlation coefficients were calculated with a Macintosh-based spreadsheet. Between-panel comparisons were also effected by way of scattergrams. The data for the various scattergrams were derived from "subject-based" data reduction. Specifically, the data from each pro-file sheet (patient) were averaged to produce a single score for that patient from each rater group. This method,

TABLE 8 Data Structure: Panel Composition*

Treatment Evaluated Panel†

Extraction Non-Extraction

1 2 Ortho . . . 67 Black

1 2 Lay . . . 32 1 2 Ortho . . . 27 White 1 2 Lay . . . 30

* Each datum is the average of a panelist's scores for either the 30 extraction or 30 non-extraction treatment outcomes.

† Ortho = Orthodontist, Lay = Layperson

therefore, preserved the rating of each individual profile comparison, thereby allowing an analysis of how each group, on average, evaluated the treatment of each patient. Thus, the data for this portion of the analysis took the structure shown in Table 9.

Finally, for purposes of discussion, the subject-based VAS means were also used to examine how each group evaluated the profile changes as a function of initial lower lip pro-trusion (lower lip to -E plane) or lower lip movement during treatment. The values for lower lip protrusion and lower lip movement were taken from Hagler's (1995) analysis of the UMAARS sample. The resulting regression lines were created by the Macintosh-based spreadsheet.

TABLE 9 Data Structure: Regression*

Treatment

Panelists Extraction Non-Extraction

1, 2, ..., 30 1, 2, ..., 30

Orthodontist Black

Layperson

Orthodontist White

Layperson

* Each datum is the average score given to a profile sheet by each panel.  

RETURN TO INDEX

RESULTS

Table 10 and Fig. 6 summarize the mean VAS scores and standard deviations for the ex-patients and all four panels: Black orthodontists (BO), White orthodontists (WO), Black laypersons (BL), and White laypersons (WL). It can be seen that the extraction subjects scored their own profiles higher than did the non-extraction ex-patients. This difference, however, was not statistically significant (p<.01). In contrast, the panels had much greater statistical power (the data from just one panelist was the equivalent of that obtained from all 60 ex-patients) and, perhaps as a result, their evaluations showed a number of statistically significant main effects and interactions.

Main effects are summarized in Table 11 and Fig 7. All scores are positive (on a scale from -100 to +100), indicating that, on average, the post-treatment profiles were preferred by all panels and groups, regardless of whether or not first premolars were extracted. It also may be seen that the extraction profiles were preferred to the non-extraction pro-files by all panels and groups.

Table 10 Mean VAS Scores and Standard Deviations

Extraction Non-Extraction Group

Mean S.D. Mean S.D.

Ex-Patients 39.56 62.56 16.22 63.21 BO 20.07 13.67 11.82 9.25 WO 28.13 10.79 13.74 6.39 BL 9.15 12.66 2.27 11.74 WL 23.02 11.14 4.81 9.06

Figure 6. The graphic depiction of the mean VAS scores for ex-patients and panels. Note that the extraction profile scores are higher for each group.

TABLE 11 Mean VAS Scores for Major Groups

Group Extraction Non-Extraction Combined

All Blacks 16.54 8.73 12.64 All Whites 25.44 9.04 17.24 All Orthodontists 22.39 12.37 10.23 All Laypersons 15.86 3.50 9.68

Figure 7. Mean VAS scores by treatment and groups: or-thodontists and laypersons. Note that the extraction scores are higher for each grouping.

The repeated-measures ANOVA (Table 12) indicates that there were significant differences between races, levels of education, and treatment strategy; however, there also was significant interaction between extraction and race. Thus, education stands as the only significant, interpretable main effect. As may be seen in Table 13, orthodontists are likely to be more approving than laypersons of the esthetic impact of orthodontic treatment.

Table 14 and Fig. 8 summarize the interaction between race and extraction. If there had been no interaction, the lines in Fig 8 would be parallel, and the scores given by Whites would always be higher than those given by Blacks. The lines here, however, converge within non-extraction treatment: Blacks and Whites are both likely to give similar low positive scores. On the extraction side, however, the scores are likely to be significantly higher for both groups, more so if the observer is White. Table 15 demonstrates statistically that the extraction scores were significantly higher than the non-extraction scores, as well as the difference between Whites and Blacks within the extraction treatment.

In summary, there are differences in the way orthodontists and laypersons evaluate the esthetic impact of orthodontic treatment in a sample of similarly-susceptible African-American orthodontic patients. Orthodontists are likely to give both treatments higher ratings than are laypersons.

TABLE 12 Repeated-measures ANOVA for Panels

Source DF SSQ MSQ F p

Race 1 2986.07 2986.07 19.47 0.00 Education 1 5110.18 5110.18 33.33 0.00 Race x Education 1 177.81 177.81 1.16 0.28

Error (Between) 152 23306.25 153.33

Extraction 1 9769.51 9769.51 106.27 0.00 Extraction x Race 1 1307.81 1307.51 14.23 0.00 Ext. x Edu. 1 25.71 25.71 0.28 0.60 Ext. x Race x Edu. 1 115.42 115.42 1.26 0.26

Error (Within) 152 13973.35 91.93

Table 13 Mean VAS Scores by Orthodontists and Laypersons

Measurement Orthodontists Laypersons

Profile Scores 17.36 9.68

TABLE 14 Race and Treatment

Treatment Race

Non-Extraction Extraction

African American 8.73 16.54

European-American 9.04 25.44

TABLE 15 1-Way ANOVA Test of Simple Main Effects

Comparison F Sig.

Between Races Within Non-Extraction 0.04 ns Between Races Within Extraction 31.40 p<.01 Between Treatment Within Blacks 32.15 p<.01 Between Treatment Within Whites 82.44 p<.01

Figure 8. Mean VAS scores by treatment and groups: Af-rican Americans and European-Americans. Note that the ex-traction scores are higher for each grouping.

Further, there is significant interaction between race and extraction: if premolars have been extracted, Whites give a significantly higher rating to the outcome of treatment.

The differences between panels lie in how strongly some favor the effects of orthodontics; however, a profile seen as appealing or unappealing by one panel is likely to be viewed more or less similarly by the other panels, regardless of race or education. Similarities in preferences are manifest in a high correlation between the scores of pairs of panels. Figs. 9 and 10 are regressions that illustrate the covariance of the overall scores for Blacks and Whites, orthodontists and laypersons. Again, these regressions show that the vari-ous groups share some general tendencies in their prefer-ences. The differences, as indicated earlier, are in the strengths of those preferences.

The subject-based correlation matrices for the average scores given by the various panels to the 60 patients are shown in Tables 16 and 17. Table 18 shows the correlation matrix for the overall (extraction and non-extraction) scores for each panel, and a graphic representation of this compari-son between can be seen in the scattergrams, Figs. 11-16.

Subject-based regressions were also utilized to examine the perceived esthetic impact of extraction and non-extraction treatment as functions of initial lower lip pro-trusion (Figs. 17-24) and lower lip movement during treatment (Figs. 25-32). The regression lines indicate that

Figure 9. Scattergram: Average scores of Blacks and Whites for 60 orthodontic patients. Note in this and other scattergrams that there is a high correlation between groups, but not a 1:1 ratio. The groups respond similarly, but with the Whites consistently giving scores about 10 points higher than Blacks. These similarities and differences are also evident in other scattergrams that compare Blacks and Whites. Figure 10. Scattergram: Average scores of orthodon-tists and laypersons for 60 orthodontic patients. Here the orthodontists give scores that are similar, but consistently about 10 points higher than layperson. This general pattern also is evident in other scattergrams that compare orthodon-tists and laypersons.

TABLE 16 Correlation Matrix: Extraction Patients

BL BO WL WO

BL 1.00 ·· ·· ·· BO 0.92 1.00 ·· ·· WL 0.90 0.93 1.00 ·· WO 0.86 0.91 0.93 1.00

TABLE 17 Correlation Matrix: Non-Extraction Patients

BL BO WL WO

BL 1.00 ·· ·· ·· BO 0.88 1.00 ·· ·· WL 0.93 0.90 1.00 ·· WO 0.83 0.88 0.87 1.00

TABLE 18 Correlation Matrix: All Patients BL BO WL WO

BL 1.00 ·· ·· ·· BO 0.91 1.00 ·· ·· WL 0.86 0.80 1.00 ·· WO 0.91 0.90 0.88 1.00

Figure 11. Scattergram: Average scores of White ortho-dontists and White laypersons for 60 orthodontic patients.

Figure 12. Scattergram: Average scores of White ortho-dontists and Black orthodontists for 60 orthodontic patients.

Figure 13. Scattergram: Average scores of White ortho-dontists and Black laypersons for 60 orthodontic patients.

Figure 14. Scattergram: Average scores of White laypersons and Black laypersons for 60 orthodontic patients.

Figure 15. Scattergram: Average scores of White laypersons and Black orthodontists for 60 orthodontic pa-tients.

Figure 16. Scattergram: Average scores of Black ortho-dontists and Black laypersons for 60 orthodontic patients.

Figure 17. Linear Regression: African-American VAS ratings as a function of initial lip protrusion for 60 ortho-dontic patients. The extraction regression line crosses the non-extraction line at about 4 mm of lower lip protrusion. The trends are similar in other regressions depicting VAS ratings as a function of lip protrusion.

Figure 18. Linear Regression: European-American VAS ratings as a function of initial lip protrusion for 60 ortho-dontic patients. Here, extraction began to have a better es-thetic impact for patients who presented with more than 1.5 mm of lip protrusion.

Figure 19. Linear Regression: Layperson VAS ratings as a function of initial lip protrusion for 60 orthodontic pa-tients. At about 2.5 mm of lip protrusion, the extraction scores began to exceed the non-extraction scores.

Figure 20. Linear Regression: Orthodontist VAS ratings as a function of initial lip protrusion for 60 orthodontic patients. Extraction led to higher scores for patients who presented with more than 3.5 mm of lip protrusion.

Figure 21. Linear Regression: African-American layper-son VAS ratings as a function of initial lip protrusion for 60 orthodontic patients. The extraction regression line crosses the non-extraction line at about 2.5 mm of lower lip protrusion.

Figure 22. Linear Regression: European-American layperson VAS ratings as a function of initial lip protrusion for 60 orthodontic patients. The regression line crosses the non-extraction line at about 2.5 mm of lower lip protrusion.

Figure 23. Linear Regression: African-American ortho-dontist VAS ratings as a function of initial lip protrusion for 60 orthodontic patients. The extraction regression line crosses the non-extraction line at about 4 mm of initial lower lip protrusion.

Figure 24. Linear Regression: European-American ortho-dontist VAS ratings as a function of initial lip protrusion for 60 orthodontic patients. The extraction regression line crosses the non-extraction line at 2.5 mm of initial lower lip protrusion.

Figure 25. Linear Regression: African-American VAS ratings as a function of lower lip movement for 60 orthodontic patients. The extraction regression line indicates that the perceived beneficial esthetic impact is greatest in ex-traction patients who had the greatest amount of lower lip retraction. Conversely, it is apparent that lip retraction has a negative impact in patients who started with relatively satisfactory profiles.

Figure 26. Linear Regression: European-American VAS ratings as a function of lower lip movement for 60 orthodontic patients. The perceived benefits of non-extraction are negative if the lower lip actually retracts more than 1.5 mm.

Figure 27. Linear Regression: Layperson VAS ratings as a function of lower lip movement for 60 orthodontic patients. The perceived impact of extraction is negative if the lips are moved forward; the perceived impact of non-extraction is negative if the lower lip is retracted.

Figure 28. Linear Regression: Orthodontist VAS ratings as a function of lower lip movement for 60 orthodontic pa-tients. The perceived impact of non-extraction is negative if the lower lip retracts more than about 2 mm.

Figure 29. Linear Regression: African-American layper-son VAS ratings as a function of lower lip movement for 60 orthodontic patients. The regression line for extraction in-dicates that there is a perceived negative impact if the lower lip protrudes more than about 3 mm. The perception is negative for non-extraction if the lip is retracted.

Figure 30. Linear Regression: European-American layperson VAS ratings as a function of lower lip movement for 60 orthodontic patients. Non-extraction is perceived as negative if the lips retract more than approximately 1 mm.

Figure 31. Linear Regression: African-American ortho-dontist VAS ratings as a function of lower lip movement for 60 orthodontic patients. For non-extraction patients, the esthetic impact is perceived as negative if the lips retract more than about 1 mm.

Figure 32. Linear Regression: European-American ortho-dontist VAS ratings as a function of lower lip movement for 60 orthodontic patients. Non-extraction is perceived as negative is the lower lip retracts more than 3 mm.

The patients who presented with the most initial lower lip protrusion (as judged by the position of the lower lip to the E plane) received the greatest perceived benefit from extrac-tion treatment. On the other hand, the regressions show a relatively flat regression line for non-extraction patients. The perceived esthetic impact of extraction treatment was about the same (low positive), regardless of initial lower lip protrusion.

The regressions depicting VAS ratings as a function of initial lower lip protrusion show a regression line having a negative slope for extraction patients and a positive slope for non-extraction patients. In Fig. 29, for example, the perceived esthetic benefit of extraction increase as a func-tion of lip retraction. On the other hand, the perceived es-thetic benefit of non-extraction treatment did not diminish as the amount of protrusive lip movement became greater.  

RETURN TO INDEX

DISCUSSION

This study compares the esthetic impact of extraction and non-extraction treatment on the profiles of a sample of similarly-susceptible African-American ex-patients. This es-thetic impact is studied both from the perspective of the pa-tients, themselves, and from the perspectives of panels of African- and European-American orthodontists and laypersons.

Methodology
University of Michigan African- American Recall Sample

The extraction and non-extraction groups of the UMAARS sample were intended to be similar subgroups of a larger par-ent sample. The relative rarity of non-extraction patients in the parent sample, however, dictated that the extraction group be chosen to match the members of the non-extraction group. In the end, their discriminant scores overlapped, and the faces, on average, appear sufficiently similar to support a relatively unbiased comparison. The present study thus benefits from the considerable size and morphological compa-rability of the two sub-groups of the UMAARS sample; it thus should permit a valid test of alternative treatments and per-ceptions.

Profile Questionnaire

When the various judges were asked to rate the esthetic appeal of faces, it was not always possible to determine just what characteristics of the profiles influenced their prefer-ences. The silhouettes employed in the present study offer the advantage of eliminating extraneous distracters such as hair texture, complexion, and facial hair. There remains, however, a possibility that judges might have been influenced simply by how artistically the profiles were drawn, or whether the patients held their lips together or apart when the cephalograms were taken. In the end, the judges were left to evaluate the profiles by their own terms, and then to characterize the strength of those preferences on the VAS. "Generalizability" of the UMAARS Sample to the African-American Community

Previous studies have presented normative cephalometric values for African Americans (e.g., Cotton and associates, 1951; Altemus, 1960; Drummond, 1968; Alexander and Hitchcock, 1978; Kowalski and associates, 1974; Fonseca and Klein, 1978; and Richardson, 1991), have proposed ratios of incisor to lip retraction in African Americans (e.g., Garner, 1974; Park and associates, 1989; and Diels and associates, 1995), and have analyzed the esthetic impact of treatment in African Ameri-cans (e.g., DeLoach, 1978; Thomas, 1970; Farrow and associ-ates, 1993; and Polk and associates, 1995). Although fewer in number than those dealing with European-American patients, previous studies of African Americans are reasonably consis-tent in their findings and serve as a foundation for the pre-sent consideration of esthetic changes in African-American patients. Specifically, facial form and treatment results seen in the ex-patients examined here seem consistent with that portion of the literature describing cephalometric norms and incisor and lip retraction in African-American subjects. This conformity can be inferred from Hagler's (1995) descrip-tion of the UMAARS sample and from the normative cephalomet-ric values for African Americans contained in Richardson's atlas (1991). Hagler also characterized in great detail how extraction and non-extraction orthodontic treatment had physically affected this sample. These changes, in effect, constituted the stimuli to which the various panels re-sponded.

Hagler's (1995) analysis of the UMAARS sample shows that, on average, the post-treatment difference between the profiles of the two groups is a matter of only a few millime-ters. Hagler suggested that the first 1-2 millimeters of in-cisor retraction served to counteract the effect of growth and/or lip-strain relief before lip retraction could occur. Thus, when the upper incisors of the extraction group were retracted an average of just over 4 mm, that movement pro-duced a reduction in lip-procumbancy of about a millimeter. The result was about a 2 mm difference between the extraction group and the non-extraction group (which became a bit fuller after treatment). Although small, this difference is clini-cally significant and led to an esthetically discernible im-pact.

The profile reduction from extraction in the present sample also is in general agreement with other studies show-ing that incisor retraction will facilitate lip retraction in African-American patients (Garner, 1974; Park, Kudlick, and Abrahamian, 1989; Diels et al., 1985) much as it does in European-American patients (Hershey, 1972; Roos, 1977; Rudee, 1964; Wisth, 1974). Similarly, comparative studies on Euro-pean-American borderline patients by Paquette, Beattie, and Johnston (1992) and Luppanapornlarp and Johnston (1993) also showed a roughly 2 mm difference between extraction and non-extraction patients.

The physical characteristics and orthodontic changes af-fecting the UMAARS sample are reviewed here because it is im-portant to understand that the observers in this study were not just looking at hypothetical esthetic consequences of or-thodontic treatment. Instead, they evaluated real changes in real patients and thus generated data that can be applied broadly to the contemporary African-American patient popula-tion. The results of treatment made an esthetic impression on the observers surveyed here. Given an understanding of these effects, practitioners can be guided in the development of individualized treatment plans. Indeed, esthetic percep-tion may well be the most important component of the extrac-tion decision.

When Hagler looked at long-term retention, he found that extraction treatment did not lead to greater irregularity in retention than did non-extraction treatment. Nor did bicus-pid extraction contribute to temporomandibular dysfunction (TMD)--there was no increased prevalence in the extraction patients. More recently, Reed (1996) found that extractions did not lead to increased periodontal disease in the UMAARS sample. Thus, concerns about irregularity in retention, TMD, or periodontal health are not significant to the extraction decision in similarly-susceptible African-American patients. This leaves facial appearance, in the end, as the key consid-eration. In this context, the present results are clear: for similarly-susceptible African-American patients, extrac-tion led to esthetic results that, on average, were preferred by all observers, regardless of race or education. The strength of this general preference, however, was a function of both background and race.

Race and Esthetic Perception

Only a mere handful of papers have examined the possi-bility that Blacks and Whites in America may look differently at the esthetic appeal of facial form. What Whites may find attractive has been inferred from Caucasian normative values, such as the ten measures of the Downs Analysis (1948), or from esthetic studies such as that of Peck and Peck (1970). The studies specifically addressing African-American prefer-ences, from DeLoach (1978) to Farrow, Zarrinnia, and Azizi (1993), consistently demonstrate that African Americans gen-erally are inclined to prefer profiles that may be straighter than can be inferred from Black normative values, although not necessarily as straight as would be called for by Cauca-sian norms. The present study is in general agreement with this point of view.

The straighter profiles created by the extraction treat-ments generally were better received than the more protrusive profiles that resulted from the non-extraction treatments. It cannot be concluded, however, that, as suggested by Martin (1964), the participants preferred the straighter profiles because they were more "White-looking." Simply, it demon-strates that protrusive profiles made less protrusive by ex-tractions were more appealing than protrusive profiles that often were made more protrusive by expansion. The present study also demonstrates that this preference is common, both to African Americans and European-Americans, and to orthodon-tists and laypersons. Differences in race and education do not lead people to greatly conflicting views on what is ap-pealing and what is not. The appeal of treated faces over non-treated faces and extraction profiles over non-extraction profiles is similar for all groups. What is different, how-ever, is how strongly some people feel about what they see.

A broad overview of the general preferences of each group can be seen in Fig. 6: all of the groups tended to like treated profiles better than non-treated profiles, ex-traction more so than non-extraction. Note also that the ex-patients' preferences were similar to those of the panels. The patients' evaluations, however, lacked statistical power: all of their data were roughly equivalent to that of only one panelist. Accordingly, it is the panelists' data that gener-ated the statistically significant outcomes.

The ratings of the various panels are highly correlated, the coefficients ranging from 0.86 to as high as 0.95. For instance, if a White orthodontist gave a profile a high posi-tive score, then a Black layperson was likely to give the same profile a high score as well (Fig. 11). High correla-tion, however does not imply a 1:1 relationship. From the data shown in Fig. 11, it can be seen that, although two groups may assign a profile scores that vary together, the scores from the White orthodontists are likely to be about 10 points higher throughout the range of scores. Thus, the scattergrams illustrate the point that different groups of people may think along the same lines when evaluating esthet-ics; however, the absolute strength of their preferences are influenced by race, extraction, and orthodontic training.

Influence of Orthodontic Training

This investigation agrees, in part, with many studies reporting that orthodontists and layperson have similar es-thetic preferences. Cox and van der Linden (1971), Dunlevy and associates (1978); Farrow, Zarrinnia, and Azizi (1993); and Oynick (1988) all indicated that orthodontists and laypersons appear to have a common sense of facial esthetics. Each study gave judges from different backgrounds the oppor-tunity to choose between different photographs or drawings that represented different treatment effects; in each in-stance, orthodontists and laypersons chose similarly.

Such was the outcome in the present study. Orthodon-tists and laypersons both liked treated profiles better than non-treated profiles; extraction better than non-extraction. Despite that, closer examination revealed that there were marked differences between the groups: the orthodontists gave significantly higher ratings. Thus, educational differ-ences turned out to be the only interpretable, significant main effect.

A similar observation was made by Spadafore (1995). He, too, found that orthodontists and laypersons agreed in their overall choices; and he, too, noted that orthodontists gave significantly different scores to specific types of treat-ment. Spadafore's study and the present study differ mark-edly with respect to the nature of these differences, how-ever.

Spadafore's observers (orthodontists, general dentists, oral surgeons, Black laypersons, and White laypersons), like the observers in the present study, were in general agreement that facial appearance had been improved by orthodontic treatment, whether or not premolars had been extracted. His observers, however, tended to be more pleased with the re-sults of non-extraction treatment. In contrast to the pres-ent study, the somewhat fuller non-extraction profiles of Spadafore's sample were deemed to be the most improved. What is the origin of the difference?

The observers in Spadafore's study knew that they were evaluating Caucasian profiles, just as the observers in the present study knew that they were evaluating Black profiles. As Peck and Peck (1970) have indicated, the general public may prefer White profiles that are more protrusive than aver-age and Black profiles that are less protrusive than average. Thus, in Spadafore's study, the panelists were least im-pressed with flat profiles that became flatter. Conversely, the observers in the present study were least impressed with full profiles that became fuller. Thus, it is easy to ex-plain the differences between the present study of Blacks and Spadafore's evaluation of Whites: the judges gave different scores to the same types of profile changes because their evaluations were rendered in light of the nature of the sub-jects they were evaluating. Interaction Between Race and Extraction In addition to an educational effect, there was a sig-nificant interaction between race and extraction. Given non-extraction treatment, Blacks and Whites, alike, viewed the outcome as only a modest improvement. In contrast, when pre-molars were extracted, both Blacks and Whites tended to award higher scores, those of the White observers being signifi-cantly higher than those of the Blacks.

Spadafore's (1995) study of Whites also showed an ex-traction interaction; however, in his investigation it was an interaction between education and extraction. The largest difference between extraction and non-extraction scores was among orthodontists, who, like the other panels of judges, preferred the non-extraction profiles and did so to a greater degree than any other group of observers. The interaction between education and extraction that caused orthodontists to be less approving of extraction results may reflect a contem-porary "backlash" against the supposed negative impact of bi-cuspid extraction. What is most apparent, however, is that Americans do not necessarily judge facial appearance by a single standard, but rather they judge each person within the context of the judge's background and the subject's race. These factors--race and education--interact on occasion to produce dramatic instances of disagreement.

Disagreement
When orthodontists and laypersons or African Americans and European-Americans offer differing opinions about facial esthetics, one would expect a visually discernible anatomical basis for this difference of opinion. The types of anatomi-cal differences that prompt such disagreement can perhaps be understood by looking at several of the profiles that gener-ated the greatest between-groups disagreement. Subject #2 (Appendix I), for instance, prompted the most disagreement (as judged by the largest numerical differences in scores) between orthodontists and laypersons. In this instance, both groups chose the fuller, pretreatment profile (choice A) over the flatter, post-treatment profile (choice B). Apparently, orthodontists and laypersons were both unimpressed with how much the profile appeared to flatten (an irony, given that this patient was treated non-extraction), so they chose the fuller, pre-treatment face. Their similar choice, however, was not uniform in strength. Orthodontists were less dis-pleased than laypersons with the flattening effect of treat-ment, judging by the average scores they gave (the orthodon-tists, -1.4; laypersons, -32.28). In other words, even when treatment had a negative effect, the orthodontists were gen-erally still more pleased than laypersons with the esthetic impact of treatment, and this is an illustration of the gen-eral findings of this study.

Similarly, Subject #10 (Appendix I) illustrates the visually discernible anatomical differences that can prompt disagreement between African Americans and European-Americans. For this extraction patient, African Americans preferred the fuller pre-treatment profile (B) and gave a mean score of -10.45. European-Americans, on the other hand, preferred the flatter post-treatment profile (A) and gave a mean score of 29.17. This difference in opinion illustrates the general finding that African Americans may like treatment that makes faces flatter, but not necessarily as flat as those preferred by European-Americans. It is likely that the African Americans examining the silhouettes for Subject #10 thought that choice A was simply too flat. It did not look like a normal, attractive African American. On the other hand, it is just as likely that the European-Americans were more receptive to the profile that more closely resembles a normal White face even though they knew they were evaluating Black profiles. They chose what was most familiar to them. Subject #10, therefore, demonstrates how the interaction be-tween race and extraction can influence the opinion of ob-servers.

Indicators for Esthetic Improvement
As suggested by Peck and Peck (1970), orthodontists may be most comfortable when they can translate the "gray areas" of esthetic improvement into more concrete scientific lan-guage. Such language can be used to form concrete indicators for use in treatment planning. With help from such indica-tors, the potential esthetic benefits of extraction or non-extraction treatment can be better anticipated. In the pres-ent study, Figs. 17-32 demonstrate perceived esthetic im-provement as functions of initial lower lip protrusion and lower lip movement during treatment. Clinically useful guidelines perhaps can be inferred from these data.

Fig. 17 depicts the mean VAS scores for Blacks plotted as a function of lower lip protrusion: the fullest profiles prior to treatment benefited the most from extraction treat-ment. In contrast, the regression line for the non-extraction patients is essentially flat, indicating that es-thetic improvement was small and independent of initial lower lip protrusion. What is important about the relationship be-tween the two regression lines is the point at which they be-gin to diverge. Beyond that point, one form of treatment be-gins to be seen as more esthetically beneficial than the other. In Fig. 17, the lines cross and begin to diverge at about 4 mm of initial lip protrusion (relative to the E plane). This divergence means that the African-American pan-elists viewed extraction as more esthetically beneficial for Black patients if they presented with more than 4 mm of lower lip protrusion. Patients who present with less than 4 mm of protrusion thus may be better served by non-extraction treat-ment.

European-Americans, on the other hand, apparently be-lieve that extraction treatment would benefit African-American patients who present with less protrusion. In Fig. 18 the extraction and non-extraction regression lines cross at an initial protrusion of only about 1.5 mm. It is at this point of divergence that Whites view extraction as more es-thetically beneficial than non-extraction. Thus, in viewing African-American profiles, the White panelists were willing to accept a somewhat fuller profile than dictated by Cauca-sian norms, but still not one as full as the Black panelists preferred. Clinical application of this information, there-fore, depends on whose esthetic preferences one is trying to please. If one is treating African-American patients, it is clear that a somewhat fuller profile is more acceptable to Blacks than to Whites.

In Figs. 25-32, esthetic improvement was plotted as a function of lower lip retraction. As such, the regression lines for extraction treatment indicate that esthetic im-provement was greater as lower lip retraction increased. It is likely that this difference is an indication that judges were most impressed by dramatic changes in initially procum-bant patients. If bicuspids were extracted, it is probable that profile flattening was thought to be beneficial, thus lip retraction would be viewed as beneficial. Concurrently, greater lip retraction did not lead to greater esthetic im-provement in non-extraction patients. For these patients, profile flattening is usually not the goal, and it would ap-pear that doing so is not viewed as beneficial.

Envoi
It is important to consider carefully how best to util-ize in daily orthodontic practice the information presented here. Orthodontists are likely to think more highly of the results of treatment than are patients. It is also likely that, if a patient is African American (the patient group currently under consideration) and the orthodontist is Euro-pean-American (as are the majority of American orthodon-tists), the patient and orthodontist may view differently the results of treatment. Yet, the existence of perceived dif-ferences must be leavened by an understanding that, in simi-larly-susceptible patients, both parties are likely to favor the results of orthodontic treatment, extraction more so than non-extraction. Despite these general trends, there is suf-ficient individual variation to warrant involving the patient in a discussion of probable esthetic results before treatment has begun, thereby allowing patients to appreciate how dif-ferent treatment options are likely to change their appear-ance, and orthodontists to understand how the patient might respond to such changes. Presumably, this approach would serve to avoid conflicting expectations or misunderstandings between the orthodontist and patient.

One possible source of misunderstanding between the pa-tient and practitioner is the assignment of "racial owner-ship," as Martin (1964) does, to profile descriptions of "flatter" or "fuller." The term "racial ownership" is used here to denote a situation in which one might describe a par-ticular type of profile, such as one that is full or flat, as belonging solely to a particular racial group. For example, were a practitioner to tell an African-American patient that extraction treatment would flatten his or her "Black-looking" profile and, thus, make it look "more White," this would be an assignment of "racial ownership" to a profile. The error in this approach is that, given the phenotypic variety within the Black race (Richardson, 1980), this patient will probably still look like a "normal" African American, even with some flattening of the profile. Martin (1964) did not see this, and others, as demonstrated in the studies presented in the Review of the Literature, have unfortunately followed his ex-ample.

More importantly, many Blacks may be taken aback by such a presumption in an era of increased African-American cul-tural awareness and self-determination. Therefore, it is ad-visable not to assign racial identification to esthetic pref-erences that all Americans seem to share: a desire to move away from the esthetic "extremes" of the distribution. Cer-tainly, such a practice would protect the practitioner from offending the cultural sensibilities of any prospective pa-tient in a discussion of esthetic improvement. African Americans may like faces that are straighter than the norma-tive values; however, that does not mean they desire faces that do not "look Black."  

RETURN TO INDEX

SUMMARY AND CONCLUSIONS

Despite an increasing demand for services by African-American patients, few studies have examined either the es-thetic preferences of African Americans or the esthetic im-pact of contemporary treatment on African-American faces. Given that extraction orthodontic treatment is commonly util-ized in the treatment of African-American patients, and given that this treatment is likely to modify facial traits that may be seen as normal and characteristic, it is important to consider how the patients themselves, and others, will view such changes. It is also important to determine whether or not such esthetic changes are viewed differently by different groups of people. Hopefully, this information will be useful to practitioners in treating patients from backgrounds dif-ferent than their own.

A group of 30 extraction and 30 non-extraction African-American patients from throughout the country were recalled to the clinics and offices at which they were treated. Ac-cording to stepwise discriminant analysis, the 60 patients were similarly susceptible to either extraction or non-extraction treatment. On recall, all had been out of treat-ment at least five years. The former patients were shown a profile questionnaire depicting, in random order, their own pretreatment and post-treatment profiles. They were asked to select which one they preferred, and then asked to rate the strength of this preference on a 100 mm Visual Analog Scale. The patients were not informed that they were looking at their own profiles.

The 60 profile sheets used in the patients' self-evaluation were then assembled into booklets. These booklets were evaluated externally by panels of African-American and European-American orthodontists and laypersons.

The findings of the present study indicate that:

1. African-American patients generally are pleased with the esthetic impact of orthodontics, whether or not premo-lars were extracted as part of treatment;

2. Both African-American and European-American orthodon-tists and laypersons tend to like the esthetic impact of orthodontic treatment on similarly-susceptible African-American patients; however, there is a tendency among all groups to prefer extraction profiles;

3. Orthodontic training and experience apparently influence orthodontists to give the esthetic impact of orthodontic treatment higher ratings than do laypersons, whether or not premolars are extracted;

4. There is an interaction between race and extraction that causes European-Americans to give higher esthetic rat-ings than African Americans if premolars are extracted, but similar ratings if premolars are not extracted;

5. As pre-treatment lower lip protrusion increases, at some point (say, from 2 to 4 mm in front of the E plane) ex-traction leads to more favored esthetic results;

6. For protrusive patients (i.e., those who were treated with extraction), measures of esthetic improvement tend to increase as a function of lip retraction.

The above findings support several general conclusions with regard to the esthetic impact of extraction and non-extraction orthodontic treatment in African Americans. First, it can be concluded that, in general, the extraction of first premolars does not reduce lip procumbancy so much as to "ruin the face." In fact, extraction treatment led to an esthetic improvement that most observers--African-American and European-American orthodontists and laypersons--preferred to the non-extraction alternative.

Second, it can be concluded that African-American and European-American orthodontists and laypersons generally share a favorable view of the esthetic impact of both extrac-tion and non-extraction orthodontic treatment in African Americans: orthodontists more so than laypersons; Whites more than Blacks; extraction more than non-extraction.

These conclusions can be put to use clinically if the practitioner understands how race, orthodontic training, and extraction influence esthetic preferences. In the end, most African-American patients will still want to "look Black," even though they may prefer profiles that are straighter than African-American normative values.  

RETURN TO INDEX

Alexander TL, Hitchcock HP. Cephalometric standards for American Negro children. Am J Orthod 1978;74:298-304.

Altemus LA. A comparison of cephalofacial relationships. Angle Orthod 1960;30:233-240.

Brace, CL. The concept of "race." In: Spencer F (ed.): History of physical anthropology: An encyclopedia. New York: Garland. [In press]

Broadbent, BH. A new x-ray technique and its application to orthodontic practice. Angle Orthod 1931;1:45-66.

Cotton WN, Takano WS, Wong WMW, Wylie WL. The Downs analysis applied to three other ethnic groups. Angle Orthod 1951;21:213-220.

Cox NH, van der Linden FPGM. Facial harmony. Am J Orthod 1971;60:175-183.

DeLoach N. Soft tissue facial profile of North American Blacks, a self assessment. [Master's thesis] Detroit, Michigan: University of Detroit, 1978.

De Smit A, Dermaut L. Soft-tissue profile preferences. Am J Orthod 1984;86:67-73.

Diels RM, Kalra V, DeLoach N, Powers M, Nelson S. Changes in soft tissue profile of African-Americans following ex-traction treatment. Angle Orthod 1995;65:285-292.

Downs WB. Variations in facial relationships: Their sig-nificance in treatment and prognosis. Am J Orthod 1948;34:812-840.

Drummond RA. A determination of cephalometric norms for the Negro race. Am J Orthod 1968;54:670-682.

Dunlevy HA, White RP, Proffit WR, Turvey TA. Professional and lay judgment of facial esthetic changes following orthognathic surgery. Int J Adult Orthod Orthog Surg 1978;2:151-158.

Farrow AL, Zarrinnia K, Azizi K. Bimaxillary protrusion in black Americans--An esthetic evaluation and the treat-ment considerations. Am J Orthod Dentofac Orthop 1993;104:240-250.

Feinstein AR. Clinical epidemiology: The architecture of clinical research. Philadelphia: WB Saunders, 1985.

Fonseca RJ, Klein WD. A cephalometric evaluation of American Negro women. Am J Orthod 1978;73:152-160.

Garner LD. Soft-tissue changes concurrent with orthodontic tooth movement. Am J Orthod 1974;66:367-377.

Hagler, BL. A long-term comparison of extraction and nonex-traction treatments in equally-susceptible African-American patients. [Master's thesis] Ann Arbor, Michi-gan: University of Michigan, 1995.

Hershey HG. Incisor tooth retraction and subsequent profile change in postadolescent female patients. Am J Orthod 1972;61:45-54.

Johnston LE Jr, Paquette DE, Beattie JR, Cassidy DW, McCray JF, Killiany DM. The reduction of susceptibility bias in retrospective comparisons of alternative treatment strategies. In: Vig KD, Vig PS (eds.): Clinical re-search as the basis of clinical practice. Volume 25. Craniofacial Growth Series. Ann Arbor: Center for Hu-man Growth and Development, The University of Michigan, 1991:155-178.

Johnston LE Jr. Clinical studies in orthodontics: Taking the low road to Scotland. In: Trotman CA, McNamara JA Jr.(eds.): Orthodontic treatment: Outcome and Effec-tiveness. Monograph 30. Craniofacial Growth Series. Ann Arbor: Center for Human Growth and Development, The University of Michigan, 1995:21-41.

Kachigan SK. Statistical analysis: An interdisciplinary in-troduction to univariate and multivariate methods. New York: Radius Press, 1986.

Kramer MS. Clinical epidemiology and biostatistics: A primer for clinical investigators and decision-makers. New York: Springer-Verlag, 1988. New York: Radius Press, 1988.

Kowalski CJ, Nasjleti CE, Walker GF. Differential diagnosis of adult make black and white populations. Angle Orthod 1974;44:346-350.

Luppanapornlarp S, Johnston LE Jr. The effects of premolar extraction: A long-term comparison of outcomes in "clear-cut" extraction and nonextraction class II pa-tients. Angle Orthod 1993;63:257-272.

Lupini JJ. A long-term comparison of contemporary edgewise treatments in African-American patients: The anatomical basis of the extraction decision. [Master's thesis] Ann Arbor, Michigan: University of Michigan, 1994.

Martin JG. Racial ethnocentrism and judgment of beauty. J of Soc Psych 1964;63:59-63.

Oynick, CD. The impact of orthodontic treatment on perceived facial esthetics. [Master's thesis] Saint Louis, Mis-souri: Saint Louis University, 1988.

Paquette DE, Beattie JR, Johnston LE Jr. A long-term com-parison of nonextraction and premolar extraction edge-wise therapy in "borderline" Class II patients. Am J Orthod Dentofac Orthop 1992;102:1-14.

Park S, Kudlick EM, Abrahamian A. Vertical dimensional changes of the lips in the North American Black patient after four first-premolar extractions. Am J Orthod Den-tofac Orthop 1989;96:152-160.

Peck H, Peck S. A concept of facial esthetics. Angle Orthod 1970;40:284-318.

Polk, MS Jr, Farman AG, Yancy JA, Gholston LR, Johnson BE. Soft tissue profile: A survey of African-American pref-erence. Am J Orthod 1995;108:90-101.

Prahl-Andersen B, Boersma H, Frans PGM, Moore AW. Percep-tions of dentofacial morphology by laypersons, general dentists, and orthodontists. J Am Dent Assoc 1979;98:209-212.

Price DD, McGrath PA, Rafii A, Buckingham B. The validation of visual analogue scales as ratio scale measures for chronic and experimental pain. Pain 1983;17:45-56.

Reed PW. A long-term comparison of periodontal effects of non-extraction and premolar extraction edgewise therapy on African Americans. [Master's thesis] Ann Arbor: University of Michigan, 1996.

Rhine S. Non-metric skull racing. In: Gill GW and Rhine S (eds.): Skeletal attributes of race: Methods for fo-rensic anthropology. Albuquerque, New Mexico: Maxwell Museum, 1990;4:9-20.

Richardson ER. Racial differences in dimensional traits of the human face. Am J Orthod 1980;50:301-311.

Richardson ER. Atlas of craniofacial growth in Americans of African descent. Volume 26. Craniofacial Growth Se-ries. Ann Arbor: Center for Human Growth and Develop-ment, The University of Michigan, 1991.

Riedel, RA. An analysis of dentofacial relationships. Am J Orthod 1957;43:103-119.

Roos N. Soft-tissue profile changes in class II treatment. Am J Orthod 1977;72:165-175.

Rudee DA. Proportional profile changes concurrent with or-thodontic therapy. Am J Orthod 1964;50:421-434.

Russell DM. Directory of orthodontists of African descent. District of Columbia: Howard University, 1994.

Spadafore, CA. Long-term esthetic impact of class II, divi-sion 1 treatments in "borderline" samples. [Master's thesis] Saint Louis University, 1995.

Steiner CC. Cephalometrics in clinical practice. Angle Or-thod 1959;29:8-29.

Sushner NI. A photographic study of the soft-tissue profile of the Negro population. Am J Orthod 1977;72:373-385.

Taylor WH, Hitchcock HP. The Alabama analysis. Am J Orthod 1966;52:245-265.

Thomas RG. An evaluation of the soft-tissue facial profile the North American black [sic] woman. Am J Orthod 1979;76:84-94.

Wisth PJ. Soft tissue response to upper incisor retraction in boys. Brit J Orthod 1974;1:199-204.