- Email: [email protected]
Predicting treatment completion in a behavioral therapy program for chronic temporomandibular pain
Journal of Psychosomatic Printed in Great Britain.
Research, Vol. 30, No. 1, pp. 57-62, 1986
PREDICTING
0
TREATMENT THERAPY
CHRONIC
COMPLETION PROGRAM
0022-3999186 $3.00 + 0.00 1986 Pergamon Press Ltd.
IN A BEHAVIORAL
FOR
TEMPOROMANDIBULAR
PAIN
DONNA P. FUNCH and ELLIOT N. GALE (Received 12 December
1984; accepted in revised form
11 May 1985)
Abstract-The aim of this study was to identify factors useful in predicting whether chronic temporomandibular pain patients would complete a behavioral treatment program. Detailed clinical examination and interviews regarding demographic and social factors were given to 78 patients one week prior to treatment. They also completed a number of personality measures (depression, anxiety, locus of control) at that time and kept a pain diary during the following week to establish baseline levels. Motivation was also assessed. Patients were classified as completing (54%) or failing to complete (46%) the program. Social factors (family and generalized others’ attitudes towards the patient’s pain) were the only significant predictors of treatment completion. In a multiple regression analysis, these factors accounted for 43% of the variance. These results suggest the usefulness of including social-environmental factors when considering patient compliance. Implications for the conceptualization of social support are discussed.
(TM) pain is a musculoskeletal disorder which may result from multiple factors [ 11. Although a variety of symptoms may be involved, pain in facial muscles and joints is the most disturbing factor and, together with headaches, is the most frequent [2, 31. Recent work suggests that non-invasive behavioral therapies, notably biofeedback and relaxation therapy, are being used increasingly to treat this disorder, as well as other types of chronic pain [ 5, 61. Evaluations of these therapies suggest they are effective with various chronic pain problems [ 71. Such evaluations, however, may be affected by the fact that a substantial number of patients drop out of treatment prematurely [ 81. Despite the significance of the problem, the dropout issue has received little attention in the area of self-management of chronic pain. If factors measured prior to the initiation of therapy could be found for use in identifying potential dropouts, strategies might be developed for minimizing and controlling this problem. Earlier research identified factors such as age [ 9, 11, 121, marital status [ 91, health locus of control [ 10, 131, depression [5, 10, 11, 14, 151 and motivation [9, 10, 12, 16, 171, together with a variety of clinical variables [9, 101, as being related to outcome. This study will consider factors and the failure to complete bahavioral therapy for chronic TM pain. TEMPOROMANDIBULAR
METHOD Sample Subjects were 78 patients with chronic TM pain who volunteered to participate in an outpatient biofeedback and relaxation training program. Subjects were predominantly female (87%) and had experienced TM pain for a minimum of 2 yr (median duration of 3.5 yr; mean duration of 6.1 yr). Prior to treatment all subjects supplied a detailed medical and dental history to evaluate factors associated with the course of this disorder. In addition, they received a carefully standardized oral examination to rule out other disorders. All subjects had been treated unsuccessfully with non-behavioral procedures: 91% Requests for reprints should be sent to Donna P. Funch, Department of Social and Preventive Medicine, State University of New York at Buffalo, 2211 Main Street, Buffalo, New York 14214, U.S.A. This research was supported, in part, by NIDR Grant No. DE 04358 from the National Institutes of Health. 57
58
DONNA P. FLINCHand ELLIOTN. GALE
had received medication; 53% received equilibration; factors constant there was no fee for treatment.
and 53% had worn mouth splints. To hold financial
Procedure Patients were randomly assigned to the following treatment conditions: taped muscle-relaxation training (n = 27); masseteric area electromyographic (EMG) biofeedback (n = 30); and a combination of both (n = 21). Treatment procedures are discussed in detail elsewhere [ 9, lo] All patients attended weekly sessions and were instructed to practice relaxation at home for 20 min daily. There were no differences among the treatment groups in terms of treatment completion or any pre-treatment study factors, based on analysis of variance procedures, allowing the consideration of the subjects as one group.
Measures One week prior to treatment a diagnostic evaluation was made and all patients were interviewed for approximately 1 hr to record demographic and clinical data. Patients also completed a number of personality measures: Wallstons’ Health Locus of Control Scale [ 181; Pilowsky’s Depression Questionnaire [ 191 ; and the Taylor Manifest Anxiety Inventory [20]. During the interview, data on the patient’s perceptions of the reaction of family members during pain episodes were also noted. These open-ended responses were rated as supportive, neutral, or non-supportive by an individual with no knowledge of the patient’s status with respect to treatment completion. Supportive statements were those indicating that family members accepted the patient’s illness and helped them in their household or work tasks. Non-supportive statements were classified as those referring to family members as unsympathetic or irritable during pain episodes. Because preliminary analysis suggested this factor to be important in treatment completion, two other true-false questions which had been asked regarding general attitudes of others toward the TM pain were also included in the analyses: (1) ‘Do people feel sorry for you when you are ill’; and (2) ‘Do you get the feeling that people are not taking your illness seriously enough’. All three questions were scored in such a fashion that a higher score indicated more sympathy and support. The true-false questions were combined to make one factor representing general attitudes. A motivation component was also included. The research assistant rated each patient’s motivation during each office session using a five-point scale ranging from totally uninvolved (1) to exceptionally involved (5). The ratings were based upon a number of behavioral criteria (e.g. completion of pain diaries). The clinical psychologist responsible for treatment (E.N.G.) also rated the patient’s motivation to provide some indication of measurement reliability. Overall reliability was high (Cronbach’s alpha = 0.93). Because the interest in this study is on identifying pretreatment predictors of completion, the motivation rating for the first session was used. Pain severity was measured in two ways. The first was based on the average time spent in pain per week (duration). The second measure represented the average weekly intensity of the pain experienced based on pain ratings (three per day) using a six-point scale ranging from zero, no pain at all, to five, pain severe enough to interfere with their behavior and to be incapacitating. These ratings were averaged to obtain weekly pain scores. Only the baseline pain score, representing pain levels for the week prior to the initiation of therapy, will be considered here. This type of rating has been validated in previous studies [21, 221. Treatment completion was represented as a dichotomous variable, incomplete vs complete. A patient’s treatment was recorded as incomplete when the patient ceased to attend weekly therapy sessions and indicated no desire to continue treatment in advance of the recommended number of sessions. The actual number of sessions attended was not used as the dependent variable since the exact number of sessions required to complete therapy varied depending on patients’ ability to acquire relaxation techniques. Thus, some subjects who completed the program may have attended fewer total sessions than some program dropouts. The most rapid completion of the therapy was five sessions. There were no significant differences between dropouts who attended fewer than five sessions and those who attended five or more so these subjects were examined together. RESULTS Four types of factors were considered in relation to treatment completion: demographic; clinical; psychological; and social (Table I). Thirty-six (46%) of the subjects failed to complete treatment, attending an average of five sessions, while the remaining 42 (54%) concluded the therapy with an averge of 15 sessions. Initially, the factors were individually considered in relatidn to treatment completion using Chi square analyses for dichotomous or nominal data and the difference of means test for interval level data. Significant differences are indicated in the last column
Predicting
TABLEI.-MEANS
treatment
59
completion
ANDSTANDARDDEVIATIONSOFSTUDYFACTORSANDTHEIRASSOCIATION WITH
TREATMENT
COMPLETION
Status Incomplete (n = 36)
Factor
Complete (n = 42)
t Values*
Demographic Age
38.5 (14.5)
41.2 (12.6)
-0.90
Sex (070 female)
81% (40%)
93% (26%)
- 1.63
Marital
status
(o/o married)
82% (39%)
92% (28%)
-1.14
number
of symptoms
5.2 (1.0)
5.4 (1.2)
-0.73
16.1 (5.5) 1.9 (0.7)
13.9 (6.3) (Z)
0.49
6.1 (5.7)
(75;
0.17
50% (51%)
55% (50%)
- 0.41
41% (50%)
64% (49%)
- 1.80
19% (40%)
41% (5Ooro)
- 2.077
3.7 (0.6)
- 1.01
Clinical Total
Severity-duration
(l-20)
-intensity
(O-5)
Time with disorder
(yr)
Past treatment-night
guard
-equilibration Wakes
with pain
1.62
Psychological Motivation
3.5 (0.8)
(1-5)
Anxiety
23.1 (9.0)
22.4 (9.0)
0.35
Depression
19.2 (6.6)
18.7 (5.8)
0.31
Health locus of control (high score= external)
32.9 (8.3)
36.4 (7.5)
- 1.92
Social Family General
attitudes attitudes
(O-2) (O-2)
1.4 (0.6)
1 .o (0.6)
- 2.80$
1.1
0.6 (0.6)
2.36t
(0.7)
*t values are presented for all factors of means test. Chi-squared analyses the findings were very similar. tp
and are the result of a two-tailed difference were also performed where appropriate and
60
DONNA P. FUNCH and ELLIOTN. GALE
of Table I. None of the demographic factors were related to treatment completion. One clinical factor, waking with pain, was associated with completion; patients who indicated that they were sometimes awakened from sleep by pain were significantly more likely to complete treatment 07 = 0.04). None of the psychological factors were related to completion. The strongest associations with the completion dimension involved social factors. The major predictor of completion was the family’s reaction to the patient’s pain 0, = 0.007); the general attitude index also supported this finding (p= 0.02). Those who completed treatment reported having families who they considered to be less supportive and more irritated and upset with them during pain episodes. These individuals were also more likely to report that people were not taking their illness seriously enough and that people did not feel sorry for them when they were ill. To clarify the roles which these factors play in predicting treatment completion, a step-wise multiple regression analysis was performed using the factors suggested above (waking with pain, equilibration, locus of control, and the social factors. These factors were not significantly intercorrelated. The social factors were the only significant predictors of treatment completion, accounting for 43% of the variance in the aependent variable (F= 16.5; p
The results of this study suggest that pre-treatment information, particularly that relating to the social environment of the patient, can be useful in predicting potential dropouts. The relations between the social factors and treatment completion were much stronger than those of other factors based on more objective and standardized measures. It is encouraging that both of the social environmental factors were significant predictors of completion despite the differences in the type of questions (open-ended vs true-false) and their administration (interview vs questionnaire). The failure of demographic and psychological factors to predict completion is consistent with more general research in the area of patient compliance [23]. Clinical factors, especially disease severity, have also been included in models used to study health-related behaviors, such as the Health Belief Model [24] and the Behavioral Model [ 251. These models suggest that the more serious the health problem, the more adherence expected, although this hypothesis has generally not been supported [ 261. This was also the case in the present research. The two separate indices of severity, duration and intensity, both failed to predict completion. Because all the patients in this study had experienced this pain for several years, it is possible that they adjusted, in some fashion, to their pain level, reducing its function as a major motivator in treatment completion. One clinical factor, being awakened by pain, was related to treatment completion. This is not a function of the severity of the pain since severity was not correlated
Predicting
treatment
completion
61
with completion. One possibility is that being awakended at night might indicate a different type of pain. It is also possible that being awakened might serve to highlight the disorder to the patient [ 271. The increase in awareness might also be accompanied by a sense of increasing interference in their lives as a result of the disorder. These factors could provide this group of patients with additional reasons for remaining in treatment. The results of the regression and discriminant function analyses emphasize the major contribution of social factors. The family’s attitude toward the patient’s pain was the single best predictor of staying in treatment. Patients who perceived family members as being irritated and upset with them were more likely to stay in treatment. The perceived attitudes of others also played a significant role. These finding are in agreement with suggestions by several investigators that family and friends can create an environment that supports or fails to support the requisite behaviors [ 23, 28, 291. These results are also consistent with Parson’s concept of secondary gain [ 301, suggesting that patients may weigh the benefits they receive from the illness, such as relief from some family obligations, to the perceived benefits associated with treatment and/or recovery. In future research, it may be valuable to distinguish between several dimensions of support, such as: (1) support which reinforces behavior patterns maintaining the health problem; (2) support for treatment-seeking activities in general; and (3) support for the particualr treatment regimen selected. The presence of support obviously has different implications for compliance depending on the type being considered. The measures used in this study were limited, relating primarily to the first type. No information was available on the attitudes of family members or friends regarding the particular type of treatment the patient was receiving. More detailed pretreatment information on this aspect of support might have further increased our ability to predict who would drop out. Family attitudes regarding treatment, however, could change once treatment is initiated and this needs to be taken into consideration as well. Additional research in this area should help to clarify the role of the social environment in treatment completion and to suggest ways in which the patient’s family can be involved in the process to help facilitate treatment and reduce the number of dropouts. REFERENCES 1. RUGH JD, SOLBERC WK. Psychological implications in temporomandibular pain and dysfunction. In Temporomandibular Joint Function and Dysfunction (Edited by ZARB GA, CARLSSON GE). St. Louis: CV Mosby, 1979. Campbell CD, Loft GH, Davis H, Hart DL. TMJ symptoms and referred pain patterns. J Prosfhet Dent 1982; 47: 430-433. SHEPPARD IM, SHEPPARD SM. Characteristics of temporomandibular joint problems. J Prosthet Dent 1977; 38: 180-191. HELKIMO M. Epidemiological survey of dysfunction of the masticatory system. In Temporomandibular Joint Fun&on and Dysfunction (Edited by ZARB GA, CARLSSONGE). St. Louis: CV Mosby, 1979. BLANCHARDEB, ANDRASIK F, NEFF DF, ARENA JG, AHLES TA, JURISH SE, PALLMEYERTP, SAUNDERS NL, TEDERS SJ, BARRON KD, RODICHOK LD. Biofeedback and relaxation training with three kinds of headache: treatment effects and their prediction. J Consult C/in Psycho/ 1982; 50: 562-575. 6. TURNER JA, CHAPMAN CR. Psychological interventions for chronic pain: a critical review. I. Relaxation training and biofeedback. Pain 1982; 12: I-21. syndrome: 7. GREENE CS, LASKIN DM. Long-term evaluation of treatment for myofascial pain-dysfunction a comparative analysis. J Am Dent Assoc 1983; 107: 235-238. 8. HART JD. Failure to complete treatment for headache: a multiple regression analysis. J Consulf C/in Psycho/ 1982; 50: 781-782.
62
DONNA P. FUNCH and ELLIOT N. GALE
9. FUNCH DP, GALE EN. Biofeedback and relaxation therapy for chronic temporomandibular joint pain: predicting successful outcomes. J Consult C/in Psycho1 1984; 52: 928-935. 10. GALE EN, FUNCH DP. Factors associated with successful outcome from behavioral therapy for chronic TMJ pain. J Psychosom Res 1984; 28: 441-448. 11. GESSELAH. Electromyographic biofeedback and tricyclic antidepressants in myofascial pain-dysfunction syndrome: psychological predictors of outcome. J Am Dent Assoc 1975; 91: 1048-1052. 12. GESSEL AH, ALDERMAN MM. Management of myofascial pain dysfunction syndrome of the temporomandibular joint by tension control training. Psychosomatics 1971; 12: 302-309. 13. ZIMET GD. Locus of control and biofeedback: a review of the literature. Pert Motor Skills 1979; 49: 871-877. 14. SCHWARTZ RA, GREENE CS, LASKIN DM. Personality characteristics of patients with myofascial paindysfunction (MPD) syndrome unresponsive to conventional therapy. JDent Res 1979; 58: 1435-1439. 15. WEPMAN LJ. Biofeedback in the treatment of chronic myofascial pain dysfunction. Psychosomatics 1980; 21: 157-162. 16. BURISH TG, HENDRIX EM. Importance of relaxation instructions in the EMG biofeedback training package. J Psychosom Res 1980; 24: 137-145. 17. KLEINMANKM. The role of reinforcement and motivation in biofeedback performance. Physiol Behav 1981; 26: 921-925. 18. WALLSTON BS, WALLTSON KA, KAPLAN GD, MAIDES SA. Development and validation of the Health Locus of Control (HLC) Scale. J Consult Cfin Psycho1 1976; 44: 580-585. 19. PILOWSKY I, SPAULDING D. A method for measuring depression: validity studies on a depression questionnaire. Br J Psychiatry 1972; 121: 411-416. 20. TAYLOR JA. A personality scale of manifest anxiety. J ABn Sot Psycho1 1953; 48: 1285-290. 21. BLANCHARD EB, ANDRASIK F, NEFF DF, JURISH SE, O’KEEFE DM. Social validation of the headache diary. Behav Therup 1981; 12: 711-715. 22. GALE EN, CORAH NL. Changes in pain upon muscle palpation after treatment for TMJ syndrome. J Dent Res 1981; 60: Special Issue A: 349. 23. KIRSCHT JP, ROSENSTOCK IM. Patients’ problems in following recommendations of health experts. In Health Psychology-A Handbook (Edited by STONE GC, COHEN F, ADLER NE). San Francisco: Jossey-Bass, 1979. 24. BECKER MH, MAIMAN LA. Sociobehavioral determinants of compliance with health and medical care recommendations. Med Care 1975; 13: 10-24. 25. ANDERSEN R. A Behavioral Model ofFumilies’ Use of Health Services. Research Series No. 25, Center for Health Administration Studies: Chicago: University of Chicago Press, 1968. 26. HAYNES RB. Determinants of compliance: the disease and the mechanics of treatment. In Compliance in Health Cure (Edited by HAYNES RB, TAYLOR DW, SACKETT DL). Baltimore: Johns Hopkins University Press, 1979. 27. CLARK CT, BEEMSTERBOERP, RUGH JD. The treatment of nocturnal bruxism using contingent EMG feedback with an arousal task. Behuv Res Therup 1981; 19: 451-455. 28. BARANOWSKIT, NADER PR, DUNN K, VANDERPOOLNA. Family self-help: promoting changes in health behavior. J Communications 1982; 32: 161-172. 29. OAKES TW, WARD JR, GRAY RM, KLAUBER MR, MOODY PM. Family expectations and arthritic patient compliance to a hand-resting splint regimen. J Chron Dis 1970; 22: 757-764. 30. PARSONS T. The Social System. New York: Free Press, 1951.
Research, Vol. 30, No. 1, pp. 57-62, 1986
PREDICTING
0
TREATMENT THERAPY
CHRONIC
COMPLETION PROGRAM
0022-3999186 $3.00 + 0.00 1986 Pergamon Press Ltd.
IN A BEHAVIORAL
FOR
TEMPOROMANDIBULAR
PAIN
DONNA P. FUNCH and ELLIOT N. GALE (Received 12 December
1984; accepted in revised form
11 May 1985)
Abstract-The aim of this study was to identify factors useful in predicting whether chronic temporomandibular pain patients would complete a behavioral treatment program. Detailed clinical examination and interviews regarding demographic and social factors were given to 78 patients one week prior to treatment. They also completed a number of personality measures (depression, anxiety, locus of control) at that time and kept a pain diary during the following week to establish baseline levels. Motivation was also assessed. Patients were classified as completing (54%) or failing to complete (46%) the program. Social factors (family and generalized others’ attitudes towards the patient’s pain) were the only significant predictors of treatment completion. In a multiple regression analysis, these factors accounted for 43% of the variance. These results suggest the usefulness of including social-environmental factors when considering patient compliance. Implications for the conceptualization of social support are discussed.
(TM) pain is a musculoskeletal disorder which may result from multiple factors [ 11. Although a variety of symptoms may be involved, pain in facial muscles and joints is the most disturbing factor and, together with headaches, is the most frequent [2, 31. Recent work suggests that non-invasive behavioral therapies, notably biofeedback and relaxation therapy, are being used increasingly to treat this disorder, as well as other types of chronic pain [ 5, 61. Evaluations of these therapies suggest they are effective with various chronic pain problems [ 71. Such evaluations, however, may be affected by the fact that a substantial number of patients drop out of treatment prematurely [ 81. Despite the significance of the problem, the dropout issue has received little attention in the area of self-management of chronic pain. If factors measured prior to the initiation of therapy could be found for use in identifying potential dropouts, strategies might be developed for minimizing and controlling this problem. Earlier research identified factors such as age [ 9, 11, 121, marital status [ 91, health locus of control [ 10, 131, depression [5, 10, 11, 14, 151 and motivation [9, 10, 12, 16, 171, together with a variety of clinical variables [9, 101, as being related to outcome. This study will consider factors and the failure to complete bahavioral therapy for chronic TM pain. TEMPOROMANDIBULAR
METHOD Sample Subjects were 78 patients with chronic TM pain who volunteered to participate in an outpatient biofeedback and relaxation training program. Subjects were predominantly female (87%) and had experienced TM pain for a minimum of 2 yr (median duration of 3.5 yr; mean duration of 6.1 yr). Prior to treatment all subjects supplied a detailed medical and dental history to evaluate factors associated with the course of this disorder. In addition, they received a carefully standardized oral examination to rule out other disorders. All subjects had been treated unsuccessfully with non-behavioral procedures: 91% Requests for reprints should be sent to Donna P. Funch, Department of Social and Preventive Medicine, State University of New York at Buffalo, 2211 Main Street, Buffalo, New York 14214, U.S.A. This research was supported, in part, by NIDR Grant No. DE 04358 from the National Institutes of Health. 57
58
DONNA P. FLINCHand ELLIOTN. GALE
had received medication; 53% received equilibration; factors constant there was no fee for treatment.
and 53% had worn mouth splints. To hold financial
Procedure Patients were randomly assigned to the following treatment conditions: taped muscle-relaxation training (n = 27); masseteric area electromyographic (EMG) biofeedback (n = 30); and a combination of both (n = 21). Treatment procedures are discussed in detail elsewhere [ 9, lo] All patients attended weekly sessions and were instructed to practice relaxation at home for 20 min daily. There were no differences among the treatment groups in terms of treatment completion or any pre-treatment study factors, based on analysis of variance procedures, allowing the consideration of the subjects as one group.
Measures One week prior to treatment a diagnostic evaluation was made and all patients were interviewed for approximately 1 hr to record demographic and clinical data. Patients also completed a number of personality measures: Wallstons’ Health Locus of Control Scale [ 181; Pilowsky’s Depression Questionnaire [ 191 ; and the Taylor Manifest Anxiety Inventory [20]. During the interview, data on the patient’s perceptions of the reaction of family members during pain episodes were also noted. These open-ended responses were rated as supportive, neutral, or non-supportive by an individual with no knowledge of the patient’s status with respect to treatment completion. Supportive statements were those indicating that family members accepted the patient’s illness and helped them in their household or work tasks. Non-supportive statements were classified as those referring to family members as unsympathetic or irritable during pain episodes. Because preliminary analysis suggested this factor to be important in treatment completion, two other true-false questions which had been asked regarding general attitudes of others toward the TM pain were also included in the analyses: (1) ‘Do people feel sorry for you when you are ill’; and (2) ‘Do you get the feeling that people are not taking your illness seriously enough’. All three questions were scored in such a fashion that a higher score indicated more sympathy and support. The true-false questions were combined to make one factor representing general attitudes. A motivation component was also included. The research assistant rated each patient’s motivation during each office session using a five-point scale ranging from totally uninvolved (1) to exceptionally involved (5). The ratings were based upon a number of behavioral criteria (e.g. completion of pain diaries). The clinical psychologist responsible for treatment (E.N.G.) also rated the patient’s motivation to provide some indication of measurement reliability. Overall reliability was high (Cronbach’s alpha = 0.93). Because the interest in this study is on identifying pretreatment predictors of completion, the motivation rating for the first session was used. Pain severity was measured in two ways. The first was based on the average time spent in pain per week (duration). The second measure represented the average weekly intensity of the pain experienced based on pain ratings (three per day) using a six-point scale ranging from zero, no pain at all, to five, pain severe enough to interfere with their behavior and to be incapacitating. These ratings were averaged to obtain weekly pain scores. Only the baseline pain score, representing pain levels for the week prior to the initiation of therapy, will be considered here. This type of rating has been validated in previous studies [21, 221. Treatment completion was represented as a dichotomous variable, incomplete vs complete. A patient’s treatment was recorded as incomplete when the patient ceased to attend weekly therapy sessions and indicated no desire to continue treatment in advance of the recommended number of sessions. The actual number of sessions attended was not used as the dependent variable since the exact number of sessions required to complete therapy varied depending on patients’ ability to acquire relaxation techniques. Thus, some subjects who completed the program may have attended fewer total sessions than some program dropouts. The most rapid completion of the therapy was five sessions. There were no significant differences between dropouts who attended fewer than five sessions and those who attended five or more so these subjects were examined together. RESULTS Four types of factors were considered in relation to treatment completion: demographic; clinical; psychological; and social (Table I). Thirty-six (46%) of the subjects failed to complete treatment, attending an average of five sessions, while the remaining 42 (54%) concluded the therapy with an averge of 15 sessions. Initially, the factors were individually considered in relatidn to treatment completion using Chi square analyses for dichotomous or nominal data and the difference of means test for interval level data. Significant differences are indicated in the last column
Predicting
TABLEI.-MEANS
treatment
59
completion
ANDSTANDARDDEVIATIONSOFSTUDYFACTORSANDTHEIRASSOCIATION WITH
TREATMENT
COMPLETION
Status Incomplete (n = 36)
Factor
Complete (n = 42)
t Values*
Demographic Age
38.5 (14.5)
41.2 (12.6)
-0.90
Sex (070 female)
81% (40%)
93% (26%)
- 1.63
Marital
status
(o/o married)
82% (39%)
92% (28%)
-1.14
number
of symptoms
5.2 (1.0)
5.4 (1.2)
-0.73
16.1 (5.5) 1.9 (0.7)
13.9 (6.3) (Z)
0.49
6.1 (5.7)
(75;
0.17
50% (51%)
55% (50%)
- 0.41
41% (50%)
64% (49%)
- 1.80
19% (40%)
41% (5Ooro)
- 2.077
3.7 (0.6)
- 1.01
Clinical Total
Severity-duration
(l-20)
-intensity
(O-5)
Time with disorder
(yr)
Past treatment-night
guard
-equilibration Wakes
with pain
1.62
Psychological Motivation
3.5 (0.8)
(1-5)
Anxiety
23.1 (9.0)
22.4 (9.0)
0.35
Depression
19.2 (6.6)
18.7 (5.8)
0.31
Health locus of control (high score= external)
32.9 (8.3)
36.4 (7.5)
- 1.92
Social Family General
attitudes attitudes
(O-2) (O-2)
1.4 (0.6)
1 .o (0.6)
- 2.80$
1.1
0.6 (0.6)
2.36t
(0.7)
*t values are presented for all factors of means test. Chi-squared analyses the findings were very similar. tp
and are the result of a two-tailed difference were also performed where appropriate and
60
DONNA P. FUNCH and ELLIOTN. GALE
of Table I. None of the demographic factors were related to treatment completion. One clinical factor, waking with pain, was associated with completion; patients who indicated that they were sometimes awakened from sleep by pain were significantly more likely to complete treatment 07 = 0.04). None of the psychological factors were related to completion. The strongest associations with the completion dimension involved social factors. The major predictor of completion was the family’s reaction to the patient’s pain 0, = 0.007); the general attitude index also supported this finding (p= 0.02). Those who completed treatment reported having families who they considered to be less supportive and more irritated and upset with them during pain episodes. These individuals were also more likely to report that people were not taking their illness seriously enough and that people did not feel sorry for them when they were ill. To clarify the roles which these factors play in predicting treatment completion, a step-wise multiple regression analysis was performed using the factors suggested above (waking with pain, equilibration, locus of control, and the social factors. These factors were not significantly intercorrelated. The social factors were the only significant predictors of treatment completion, accounting for 43% of the variance in the aependent variable (F= 16.5; p
The results of this study suggest that pre-treatment information, particularly that relating to the social environment of the patient, can be useful in predicting potential dropouts. The relations between the social factors and treatment completion were much stronger than those of other factors based on more objective and standardized measures. It is encouraging that both of the social environmental factors were significant predictors of completion despite the differences in the type of questions (open-ended vs true-false) and their administration (interview vs questionnaire). The failure of demographic and psychological factors to predict completion is consistent with more general research in the area of patient compliance [23]. Clinical factors, especially disease severity, have also been included in models used to study health-related behaviors, such as the Health Belief Model [24] and the Behavioral Model [ 251. These models suggest that the more serious the health problem, the more adherence expected, although this hypothesis has generally not been supported [ 261. This was also the case in the present research. The two separate indices of severity, duration and intensity, both failed to predict completion. Because all the patients in this study had experienced this pain for several years, it is possible that they adjusted, in some fashion, to their pain level, reducing its function as a major motivator in treatment completion. One clinical factor, being awakened by pain, was related to treatment completion. This is not a function of the severity of the pain since severity was not correlated
Predicting
treatment
completion
61
with completion. One possibility is that being awakended at night might indicate a different type of pain. It is also possible that being awakened might serve to highlight the disorder to the patient [ 271. The increase in awareness might also be accompanied by a sense of increasing interference in their lives as a result of the disorder. These factors could provide this group of patients with additional reasons for remaining in treatment. The results of the regression and discriminant function analyses emphasize the major contribution of social factors. The family’s attitude toward the patient’s pain was the single best predictor of staying in treatment. Patients who perceived family members as being irritated and upset with them were more likely to stay in treatment. The perceived attitudes of others also played a significant role. These finding are in agreement with suggestions by several investigators that family and friends can create an environment that supports or fails to support the requisite behaviors [ 23, 28, 291. These results are also consistent with Parson’s concept of secondary gain [ 301, suggesting that patients may weigh the benefits they receive from the illness, such as relief from some family obligations, to the perceived benefits associated with treatment and/or recovery. In future research, it may be valuable to distinguish between several dimensions of support, such as: (1) support which reinforces behavior patterns maintaining the health problem; (2) support for treatment-seeking activities in general; and (3) support for the particualr treatment regimen selected. The presence of support obviously has different implications for compliance depending on the type being considered. The measures used in this study were limited, relating primarily to the first type. No information was available on the attitudes of family members or friends regarding the particular type of treatment the patient was receiving. More detailed pretreatment information on this aspect of support might have further increased our ability to predict who would drop out. Family attitudes regarding treatment, however, could change once treatment is initiated and this needs to be taken into consideration as well. Additional research in this area should help to clarify the role of the social environment in treatment completion and to suggest ways in which the patient’s family can be involved in the process to help facilitate treatment and reduce the number of dropouts. REFERENCES 1. RUGH JD, SOLBERC WK. Psychological implications in temporomandibular pain and dysfunction. In Temporomandibular Joint Function and Dysfunction (Edited by ZARB GA, CARLSSON GE). St. Louis: CV Mosby, 1979. Campbell CD, Loft GH, Davis H, Hart DL. TMJ symptoms and referred pain patterns. J Prosfhet Dent 1982; 47: 430-433. SHEPPARD IM, SHEPPARD SM. Characteristics of temporomandibular joint problems. J Prosthet Dent 1977; 38: 180-191. HELKIMO M. Epidemiological survey of dysfunction of the masticatory system. In Temporomandibular Joint Fun&on and Dysfunction (Edited by ZARB GA, CARLSSONGE). St. Louis: CV Mosby, 1979. BLANCHARDEB, ANDRASIK F, NEFF DF, ARENA JG, AHLES TA, JURISH SE, PALLMEYERTP, SAUNDERS NL, TEDERS SJ, BARRON KD, RODICHOK LD. Biofeedback and relaxation training with three kinds of headache: treatment effects and their prediction. J Consult C/in Psycho/ 1982; 50: 562-575. 6. TURNER JA, CHAPMAN CR. Psychological interventions for chronic pain: a critical review. I. Relaxation training and biofeedback. Pain 1982; 12: I-21. syndrome: 7. GREENE CS, LASKIN DM. Long-term evaluation of treatment for myofascial pain-dysfunction a comparative analysis. J Am Dent Assoc 1983; 107: 235-238. 8. HART JD. Failure to complete treatment for headache: a multiple regression analysis. J Consulf C/in Psycho/ 1982; 50: 781-782.
62
DONNA P. FUNCH and ELLIOT N. GALE
9. FUNCH DP, GALE EN. Biofeedback and relaxation therapy for chronic temporomandibular joint pain: predicting successful outcomes. J Consult C/in Psycho1 1984; 52: 928-935. 10. GALE EN, FUNCH DP. Factors associated with successful outcome from behavioral therapy for chronic TMJ pain. J Psychosom Res 1984; 28: 441-448. 11. GESSELAH. Electromyographic biofeedback and tricyclic antidepressants in myofascial pain-dysfunction syndrome: psychological predictors of outcome. J Am Dent Assoc 1975; 91: 1048-1052. 12. GESSEL AH, ALDERMAN MM. Management of myofascial pain dysfunction syndrome of the temporomandibular joint by tension control training. Psychosomatics 1971; 12: 302-309. 13. ZIMET GD. Locus of control and biofeedback: a review of the literature. Pert Motor Skills 1979; 49: 871-877. 14. SCHWARTZ RA, GREENE CS, LASKIN DM. Personality characteristics of patients with myofascial paindysfunction (MPD) syndrome unresponsive to conventional therapy. JDent Res 1979; 58: 1435-1439. 15. WEPMAN LJ. Biofeedback in the treatment of chronic myofascial pain dysfunction. Psychosomatics 1980; 21: 157-162. 16. BURISH TG, HENDRIX EM. Importance of relaxation instructions in the EMG biofeedback training package. J Psychosom Res 1980; 24: 137-145. 17. KLEINMANKM. The role of reinforcement and motivation in biofeedback performance. Physiol Behav 1981; 26: 921-925. 18. WALLSTON BS, WALLTSON KA, KAPLAN GD, MAIDES SA. Development and validation of the Health Locus of Control (HLC) Scale. J Consult Cfin Psycho1 1976; 44: 580-585. 19. PILOWSKY I, SPAULDING D. A method for measuring depression: validity studies on a depression questionnaire. Br J Psychiatry 1972; 121: 411-416. 20. TAYLOR JA. A personality scale of manifest anxiety. J ABn Sot Psycho1 1953; 48: 1285-290. 21. BLANCHARD EB, ANDRASIK F, NEFF DF, JURISH SE, O’KEEFE DM. Social validation of the headache diary. Behav Therup 1981; 12: 711-715. 22. GALE EN, CORAH NL. Changes in pain upon muscle palpation after treatment for TMJ syndrome. J Dent Res 1981; 60: Special Issue A: 349. 23. KIRSCHT JP, ROSENSTOCK IM. Patients’ problems in following recommendations of health experts. In Health Psychology-A Handbook (Edited by STONE GC, COHEN F, ADLER NE). San Francisco: Jossey-Bass, 1979. 24. BECKER MH, MAIMAN LA. Sociobehavioral determinants of compliance with health and medical care recommendations. Med Care 1975; 13: 10-24. 25. ANDERSEN R. A Behavioral Model ofFumilies’ Use of Health Services. Research Series No. 25, Center for Health Administration Studies: Chicago: University of Chicago Press, 1968. 26. HAYNES RB. Determinants of compliance: the disease and the mechanics of treatment. In Compliance in Health Cure (Edited by HAYNES RB, TAYLOR DW, SACKETT DL). Baltimore: Johns Hopkins University Press, 1979. 27. CLARK CT, BEEMSTERBOERP, RUGH JD. The treatment of nocturnal bruxism using contingent EMG feedback with an arousal task. Behuv Res Therup 1981; 19: 451-455. 28. BARANOWSKIT, NADER PR, DUNN K, VANDERPOOLNA. Family self-help: promoting changes in health behavior. J Communications 1982; 32: 161-172. 29. OAKES TW, WARD JR, GRAY RM, KLAUBER MR, MOODY PM. Family expectations and arthritic patient compliance to a hand-resting splint regimen. J Chron Dis 1970; 22: 757-764. 30. PARSONS T. The Social System. New York: Free Press, 1951.