ORIGINAL ARTICLE

Comparing subjective and objective measures

of headgear compliance
Annemieke Bos,a Cornelis J. Kleverlaan,b Johan Hoogstraten,c Birte Prahl-Andersen,d and Rein Kuitertd
Amsterdam, the Netherlands

Introduction: Many studies have used subjective measures to examine patient compliance during orth-
odontic treatment. Objective measurement of compliance has been confined to only a few studies that used
electronic timing devices built into removable appliances. Our aim in this study was to compare subjective
and objective methods of measuring compliance with headgear wear. It was hypothesized that orthodontists,
patients, and patients’ parents overestimate compliance and report more wearing hours than the headgear
timers indicate. Also, relationships between sex, age, treatment time, and headgear compliance were
explored. Methods: A headgear timer device and 3 questionnaires were developed to assess compliance.
The subjects were unaware that their headgear use was being measured. Results: Significant differences
between the estimates of orthodontists, patients, parents, and headgear timer scores were found. Also, there
were differences regarding age and treatment time. Conclusions: Subjective measures might result in
overestimation of compliance. This suggests that the use of an objective instrument to measure headgear
compliance should be continued in future studies. (Am J Orthod Dentofacial Orthop 2007;132:801-5)

of oral hygiene.2 These methods commonly result in

T
o achieve a successful orthodontic treatment
outcome, patients are expected to follow regi- overestimation of compliance.3 Patient and parent judg-
mens recommended by the orthodontist. Pa- ments about compliance have also been found to be
tients must practice good oral hygiene, follow dietary unreliable.4,5
restrictions, attend appointments, and wear removable Objective measurement of compliance has been
devices, such as headgear or elastics, as prescribed. confined to only a few studies that used electronic
Although patient compliance is important in all types of timing devices built into removable appliances. In
orthodontic treatment, it is particularly so with wearing 1974, the first study with a headgear timing device was
headgear, because noncompliance can lead to increased presented.6 However, this device was found to be
treatment time, increased costs, and the necessity to use inaccurate and susceptible to patients’ attempts to
a less-preferred alternative treatment.1 falsify results.7,8 In 1991, a headgear timer based on a
Many studies have examined patient compliance small quartz calendar watch was developed.9 With this
during orthodontic treatment by using subjective mea- instrument, poor correlation was found between the
sures, such as patient or parent reports or estimates by hours the patient reported wearing the headgear and the
orthodontists. The validity of these indirect measures is actual hours worn.10 In a more recent study with
questionable. Orthodontists’ ratings of compliance headgear timers, it was suggested that patients report
often are biased by visible clues, such as the appearance more headgear wear than actually occurs.11
of the appliances, the ease of patient use, or the degree Our aim in this study was to compare subjective and
objective methods of measuring compliance with head-
a
Professor, Department of Social Dentistry and Behaviorial Sciences and
Department of Orthodontics, Academic Centre for Dentistry Amsterdam,
gear. It was hypothesized that orthodontists, patients,
Amsterdam, the Netherlands. and patients’ parents overestimate compliance and re-
b
Professor, Department of Dental Material Science, Academic Centre for port more wearing hours than headgear timers would
Dentistry Amsterdam, Amsterdam, the Netherlands.
c
Professor, Department of Social Dentistry and Behaviorial Sciences, Aca-
indicate. Also, relationships between sex, age, treat-
demic Centre for Dentistry Amsterdam; Professor, Department of Psycholog- ment time, and headgear compliance were explored.
ical Methods, University of Amsterdam, Amsterdam, the Netherlands.
d
Professor, Department of Orthodontics, Academic Centre for Dentistry Am-
sterdam, Amsterdam, the Netherlands. MATERIAL AND METHODS
Reprint requests to: Annemieke Bos, Academic Centre for Dentistry Amster-
dam, Department of Social Dentistry and Behavioral Sciences, Louwesweg 1, A headgear timer device was developed for this
1066 EA Amsterdam, the Netherlands; e-mail, [email protected]. study. Essentially, the timer recorded temperature. A
Submitted, October 2005; revised and accepted, January 2006. small device, 16 mm in diameter and 6 mm high
0889-5406/$32.00
Copyright © 2007 by the American Association of Orthodontists. (Thermochron i-Button, DS1921G; Maxim Integrated
doi:10.1016/j.ajodo.2006.01.041 Products, Sunnyvale, Calif), was mounted in the center
801
802 Bos et al American Journal of Orthodontics and Dentofacial Orthopedics
December 2007

of the headgear neckstrip. At this position, there is sion of the Free University of Amsterdam, informed
optimal contact between the neckstrip and the skin. The consent was given after the research. At the follow-up
recording was done at a rate of 1 measurement per 20 appointment, about a month later, the patient received a
minutes for 29 days. Directly after the follow-up new neckstrip, without a timer. The timer was collected
appointment, the data were transferred to a personal by the researchers, and the parent (if available), patient,
computer. and orthodontic resident were requested to complete a
Three questionnaires, developed based on pub- questionnaire immediately after the follow-up appoint-
lished literature, were evaluated in pretests with pa- ment. To guarantee anonymity, each patient and parent
tients, parents, and orthodontists. The patient question- completed his or her form in a room separate from the
naire contained 9 items about headgear wear. Patients treatment area. Two research assistants asked patients
were asked whether they had worn their headgear in the and parents not to talk while completing the forms.
last month during school hours, afternoons, evenings, Meanwhile, the orthodontist completed the form in the
weekends, and nights. Also, they were asked whether treatment room.
they had worn their headgear as recommended by the
orthodontist. All items except 1 were answered on a Statistical analyses
5-point scale, ranging from 1 (totally agree) to 5 (totally
The total hours of wear time and the mean hours of
disagree). One item asked the patient to indicate how
wear per day over the 29-day period were based on the
long the headgear was worn per day by writing the
assumption that the headgear was being worn when the
number of hours.
thermistor recorded a temperature ⬎30°C. The day was
The parent questionnaire contained the same items
divided into school hours (8:00-15:00), afternoon (15:00-
as the children’s questionnaire, but the items were
18:00), evening (18:00-22:00), and night (22:00-8:00)
slightly reformulated (instead of “I have worn my
during weekdays, and day (8:00-22:00) and night
headgear in the evening after dinner,” parents were
(22:00-8:00) on weekends. Furthermore, a day started
asked “My child wore his/her headgear in the evening
at 22:00 and ended at 21:59 to avoid double counting of
after dinner”).
the total number of days the headgear was worn in the
The orthodontist questionnaire consisted of 9 items.
29-day period.
Orthodontists were asked whether they could tell if the
To compare the estimates of orthodontists, parents,
patient had been compliant by looking at his or her
and patients with the results of the headgear timer,
teeth and headgear, whether they believed the patient
Student paired-samples t tests were performed. A cor-
overestimated or underestimated the compliant behav-
relational analysis was done to examine correlations
ior, and whether they thought that patient was honest
between the different measurements.
about wearing behavior. Also, they were asked how
Student independent-samples t tests were used to
many hours the patient reported to have worn the
explore differences in headgear wear as measured by
headgear each day and to estimate how many hours a
the timers and as reported by patient, parent, and
day the patient had worn his headgear. Again, items
orthodontist with regard to sex, age, and treatment time.
were answered on a 5-point scale, ranging from 1
Patients were divided first into 2 groups based on
(totally agree) to 5 (totally disagree), except for the 2
treatment time at the assessment (the median of 8
items on the number of hours.
months was the cutoff point) and then into 2 groups
The study took place at the Department of Orth-
based on age (the median of 13 years was the cutoff
odontics at Academic Centre for Dentistry Amsterdam.
point). As recommended by Bonferroni, to reduce the
The sample comprised 56 orthodontic patients (66%
type 1 error rate, for each independent-samples t test,
female) ranging in age from 10 to 22 years (mean age,
the critical significance level was adjusted to 0.01.12
12.89 years; SD 2.16; median, 13 years), who received
Also, a correlational analysis regarding treatment time,
treatment with low-pull headgear. Some patients were
age, and headgear use was conducted.
near the end of headgear treatment, whereas others
Because of missing values, the number of respon-
were starting. The start of orthodontic treatment varied
dents in the analyses was variable.
between 1 and 38 months (mean treatment time at the
assessment, 9.61 months; SD 8.44; median, 8 months).
Every headgear patient who visited the orthodontic RESULTS
department between April and June 2005 received a Three patients did not complete the questionnaire
neckstrip in which a timer was hidden. The patients because of lack of time. Twenty-five patients were
were not informed during the study about this instru- accompanied by a parent at the follow-up appointment;
ment. In agreement with the medical ethical commis- these parents participated in the study. All patients were
American Journal of Orthodontics and Dentofacial Orthopedics Bos et al 803
Volume 132, Number 6

Table I. Results of the correlation analysis

Amount of hours
wearing time per
day as indicated by: Timer Orthodontist Patient Parent

Timer 1.00
Orthodontist 0.73† 1.00
Patient 0.50† 0.58† 1.00
Parent 0.52* 0.45 0.98† 1.00

*P ⬍.01; P ⬍.001.
†

parents, and headgear timers. Compared with the mea-

surements of the headgear timer over the 29 days (5.58
Fig 1. The Thermochron i-Button. hours a day), orthodontists estimated that patients wore
their headgear about 9.52 hours a day (SD 3.59) (t ⫽
⫺9.46; df ⫽ 54; P ⬍.001), patients reported wearing
their headgear about 11.02 hours a day (SD 3.77) (t ⫽
⫺8.98; df ⫽ 48; P ⬍.001), and parents estimated that
the headgear had been worn 11.12 hours a day (SD
3.97) (t ⫽ ⫺5.12; df ⫽ 25; P ⬍.001).
The results of the correlational analysis are pre-
sented in Table I. Correlations between the different
instruments were significant but far from perfect. Only
the estimates of parents and orthodontists did not
correlate significantly.
Orthodontists reported that patients said they
wore their headgear about 11 hours a day (mean,
11.07; SD 3.54), but they believed that the patients
Fig 2. Typical output of the headgear timers of 3
patients. The recorded time that the temperature was
had worn their headgear about 9 hours a day (mean,
more than 30°C corresponds with the time that the 9.59; SD 3.64) (t ⫽ 6.75; df ⫽ 52; P ⬍.001). On a
headgear was worn. Patient 1 (top) was the most scale from 1 (totally agree) to 5 (totally disagree),
compliant patient with a total wear time of 11.0 hours orthodontists scored 2.41 (SD 1.55) when asked
per day. Patient 2 (middle) was less compliant and wore whether they believed the patient had been compliant
the headgear only 20 days of the 29-day period, with an in wearing headgear. When asked whether they
average of 3.6 hours. Patient 3 (bottom) used the believed patients were honest about their wearing
headgear only once for 9.7 hours, giving an average of behavior, orthodontists scored 2.29 (SD 1.04).
0.3 hours for the 29 days. Patients reported they had worn their headgear
according to the orthodontist’s advice. On the item “I
monitored for 29 days. Figure 1 illustrates the Thermo- have worn my braces in the past 4 weeks just as much
chron i-Button, and Figure 2 shows the headgear timer as the orthodontist advised me to do,” patients scored
output. 2.13 (SD 1.11), and on the item “I have worn my
The patients wore their headgear approximately 17 headgear according to the advice of the orthodontist,”
days during this period (SD 9.92; range, 0-28). Over the patients scored 1.94 (SD 1.55). Also, the parents
29 days, mean wearing time was 5.58 hours (SD 4.39; believed that their children were compliant; their mean
range, 0-18.5) per day. When only the days that the scores on these 2 items were 1.72 (SD 0.98) and 1.35
headgear was actually worn were considered, mean (SD 0.69), respectively.
wearing time was 7.61 hours (SD 4.16; range, 0-19.1) Independent-samples t tests showed no differences
per day. The patients wore their headgear about 76.7% between female and male patients. However, differ-
of the time at night, 5.5% in the evening, 8.7% of the ences with respect to age and treatment time were
time on weekends, 6.9% of the time at school, and 2.2% found, as shown in Table II. Patients younger than 13
of the time in the afternoon. years wore their headgear longer than patients 13 years
Paired-samples t tests showed significant differ- or older, and patients who had been treated for less than
ences in the measurements of orthodontists, patients, 8 months wore their headgear more often and longer
804 Bos et al American Journal of Orthodontics and Dentofacial Orthopedics
December 2007

Table II. Differences between patients regarding age and treatment time
Mean SD Mean SD t P

Age ⬍13 years (n ⫽ 26) ⱖ13 years (n ⫽ 30)

Wearing time over days headgear was actually worn 9.38 3.59 6.08 4.06 ⫺3.19 .002
Treatment time ⬍8 months (n ⫽ 26) ⱖ8 months (n ⫽ 30)
Total wearing time, night 162.50 95.11 95.16 84.19 ⫺2.81 .007
Total wearing time, evening 14.62 13.70 4.30 8.45 ⫺3.44 .001
Total wearing time, weekend 24.48 36.21 5.58 6.48 ⫺2.81 .007
Wearing time over 29-day period 7.53 4.74 3.89 3.30 ⫺3.37 .001
Wearing time over days headgear was worn 9.61 3.79 5.88 3.71 ⫺3.72 .000

Wearing time reported in hours.

Table III. Results of correlational analysis regarding headgear use, age, and treatment time
1 2 3 4 5 6 7 8 9 10

1 Age 1.00
2 Time in treatment .40† 1.00
3 Total wearing time, night ⫺.17 ⫺.33* 1.00
4 Total wearing time, school ⫺.27* ⫺.23 .26 1.00
5 Total wearing time, afternoon ⫺.22 ⫺.28* .31* .66† 1.00
6 Total wearing time, evening ⫺.31* ⫺.47† .60† .40† .72† 1.00
7 Total wearing time, weekend ⫺.14 ⫺.32* .45† .37† .50† .67† 1.00
8 Total days headgear was worn ⫺.19 ⫺.30* .89† .46† .37† .55† .45† 1.00
9 Mean wearing time over days headgear was worn ⫺.31* ⫺.45† .86† .43† .51† .69† .51† .77† 1.00
10 Mean wearing time over 29-day period ⫺.25 ⫺.40† .94† .54† .58† .76† .58† .90† .90† 1.00

*Correlation is significant at 0.05 level; †correlation is significant at 0.01 level.

than patients treated for a longer time. The results of the pliance. It was found before that compliance worsens
correlational analysis confirmed these findings. Table during a prolonged treatment regimen.16,17 This finding
III shows negative correlations between headgear use, can be explained by “patient burnout.”18 Because many
age, and treatment time. esthetic problems of orthodontic patients are usually
solved in the first 9 months of treatment, patients might
DISCUSSION
then start to think that nothing remains to be done,
This study demonstrates considerable discrepancies whereas orthodontists know that more treatment time is
among the various methods used to assess headgear needed to achieve the desired occlusal results. This
compliance. As expected, orthodontists, patients, and might cause a discrepancy between patients’ and orth-
parents significantly overestimated headgear compli- odontists’ expectations and, finally, in a decrease of
ance. Parents reported the most hours, followed by compliance.18
patients and orthodontists. Patients told their orthodon-
Some limitations of this study should be mentioned.
tists that they had worn their headgear compliantly, but
First, compliance was determined over a relatively
the orthodontists believed that patients overestimated
short time. Thus, no information about the regularity or
the hours of wearing. However, the orthodontists also
continuity of wear was assessed. Second, the impact of
overestimated the amount of time the headgear was
worn. measurement reactivity was considered negligible with
Although in previous studies it was stated that regard to patients and parents, since they were unaware
females are more compliant than males,13 we found that they were taking part in the study, but it might have
no sex differences. In agreement with earlier studies, affected the behavior of orthodontists because they
age was significantly associated with patient compli- knew about the research. In a follow-up study, a
ance.4,14,15 A possible explanation for this finding is double-blind design is recommended. Furthermore, the
that preadolescent children might be more receptive sample in this study consisted of patients, parents, and
and obedient to parental influence, and thus more orthodontic residents at Academic Centre for Dentistry
responsive to instructions than adolescents. Also, treat- Amsterdam. In other settings, the results might be
ment time was significantly related to headgear com- different.
American Journal of Orthodontics and Dentofacial Orthopedics Bos et al 805
Volume 132, Number 6

CONCLUSIONS 8. Banks PA, Read MJF. An investigation into the reliability of the
timing headgear. Br J Orthod 1987;14:263-7.
Subjective measures used in compliance studies 9. Cureton SL, Regennitter F, Orbell MG. An accurate, inexpensive
might result in overestimation of compliance. This headgear timer. J Clin Orthod 1991;25:749-54.
suggests that the use of an objective instrument to 10. Cureton SL, Regennitter FJ, Yancey JM. The role of the
headgear calendar in headgear compliance. Am J Orthod Dento-
measure headgear compliance should be continued in
facial Orthop 1993;104:387-94.
future studies. 11. Cole WA. Accuracy of patient reporting as an indication of
headgear compliance. Am J Orthod Dentofacial Orthop 2002;
We thank Hanny Alwicher, Jenninke de Jong,
121:419-23.
Ashley Pimontel, Tessa Schaaper, and Lisa Snip for 12. Stevens J. Applied multivariate statistics for the social sciences.
their assistance in collecting the data. Mahwah, NJ: Lawrence Erlbaum Associates; 1996.
13. Cucalon A, Smith RJ. Relationship between compliance by
adolescent orthodontic patients and performance on psycholog-
REFERENCES ical tests. Angle Orthod 1989;60:108-14.
1. Lyons EK, Ramsay DS. Preliminary tests of a new device to 14. Cureton SL, Regennitter FJ, Yancey JM. Clinical versus quan-
monitor orthodontic headgear use. Semin Orthod 2002;8:29-34. titative assessment of headgear compliance. Am J Orthod Dento-
2. Bos A, Hoogstraten J, Prahl-Andersen B. Towards a comprehen- facial Orthop 1993;104:277-84.
sive model for the study of compliance in orthodontics. Eur 15. Richter DD, Nanda RS, Sinha PK, Smith DW. Effect of behavior
J Orthod 2005;27:296-301. modification on patient compliance in orthodontics. Angle
3. Lyons EK, Ramsay DS. A self-regulation model of patient Orthod 1998;68:123-32.
compliance in orthodontics: implications for the design of a 16. Berndsen M, Eijkman MAJ, Hoogstraten J. Compliance per-
headgear monitor. Semin Orthod 2000;6:224-30. ceived by Dutch periodontists and hygienists. J Clin Periodontol
4. Bartsch A, Witt E, Sahm G, Schneider S. Correlates to objective 1993;20:668-72.
patient compliance with removable appliance wear. Am J Orthod 17. Kahl B, Fischbach H, Schwarze CW. How to deal with the
Dentofacial Orthop 1993;104:378-86. drop-out in clinical follow-up studies: results of a long-term
5. Sahm G, Bartsch A, Witt E. Reliability of patient reports on follow-up study of orthodontically treated patients. Am J Orthod
compliance. Eur J Orthod 1990;12:438-46. Dentofacial Orthop 1995;108:415-20.
6. Northcutt M. The timing headgear. J Clin Orthod 1974;8:321-4. 18. Brezniak N, Ben-Ya’ir S. Patient burnout— behaviour of young
7. Clemmer EJ, Hayes EW. Patient cooperation in wearing orth- adults undergoing orthodontic treatment. Stress Med 1989;5:
odontic headgear. Am J Orthod 1979;75:519-24. 183-7.