Psychiatric Annals

The MMPI in Medical Practice

David Osborne, PhD

Abstract

When Hathaway and McKinley first developed the MMPI they intended it to be used with all types of patients. The test's initial name. "Medical and Psychiatric Inventory." conveyed this intention. Alter citing the large number ol patients who seek medical attention because of neurotic complaints and the large number of patients with organic disease who suffer concomitant emotional reactions, they suggested that it would be useful to have an inventory that would help distinguish between functional and organic complaints, identify a functional component in organic disease, and identify the emotional distress of medical patients.

Long interviews with medical patients were sometimes difficult on medical wards and it was hoped that the new inventory would allow the patient to report symptoms, complaints, and other personal information in a private, unbiased way- The inventory was also expected to save a considerable amount of professional time.

There was some concern at first that patients would be upset by the MMPI questions or that they would be offended by being asked to take the lest. McKinley and Hathaway reported very little such reaction from patients.1 A similar experience was reported by Swenson et al when they sampled 1.359 medical patients who had been asked by an internist to take the MMPI.-'They found that 89.2% completed the test with no difficulty. 5.8% either refused to cooperate or did not complete the test, and 5% were unable to cooperate (nonreaders. acutely ill, etc.). They concluded that patients accept the idea of completing a personality inventory as part of a medical evaluation. They also surveyed 158 physicians who had used the MMPl with all of their patients. This 14-item questionnaire produced a 70% to 85% endorsement of the test utility. The MMPl therefore appears to be accepted by both physicians and patients as part of a medical evaluation.

In the early literature, several investigators'-4 presented case studies that illustrated the use of the MMPI with medical patients but Guthrie5 was the first to study the use of the test in a medical setting systematically. After an internist had taken the history and recorded his impression, Guthrie classified the MMPI profiles according to two-point codes (a listing of the highest two clinical scales). He studied the medical records of 365 men and 739 women in an attempt to develop a summary of the most frequently appearing physical and psychologic symptoms. Summary descriptions for five MMPI code types were developed and then evaluated by using 66 patients from the practice of a different internist. The findings supported the original descriptions.

Because previous studies of the MMPI in medical patients had used either a small number of subjects or a highly selected sample, psychologists at the Mayo Clinic were reluctant to apply these results to their patient group. Using automated techniques they collected a large number of MMPIs from medical patients; for a 3-year period all patients seen by a group of internists at the Mayo Clinic were given the MMPi.6 The subjects were ambulatory outpatients. 16 years of age or older, who had sufficient reading ability to complete the MMPI. The project continued until valid MMPI profiles of 50,000 medical patients had been collected. It was found that the medical patients had higher scores than the MMPI standardization group on a number of scales. Scores on scales J (Hs), 2 (D), and 3 (Hy). which reflect somatic concern, depressed mood, and denial of emotional problems, were quite elevated relative to the standardization group. Therefore, they concluded that elevations on these scales should be interpreted with caution in a medical population. They believed that these elevations were not necessarily the…

When Hathaway and McKinley first developed the MMPI they intended it to be used with all types of patients. The test's initial name. "Medical and Psychiatric Inventory." conveyed this intention. Alter citing the large number ol patients who seek medical attention because of neurotic complaints and the large number of patients with organic disease who suffer concomitant emotional reactions, they suggested that it would be useful to have an inventory that would help distinguish between functional and organic complaints, identify a functional component in organic disease, and identify the emotional distress of medical patients.

Long interviews with medical patients were sometimes difficult on medical wards and it was hoped that the new inventory would allow the patient to report symptoms, complaints, and other personal information in a private, unbiased way- The inventory was also expected to save a considerable amount of professional time.

There was some concern at first that patients would be upset by the MMPI questions or that they would be offended by being asked to take the lest. McKinley and Hathaway reported very little such reaction from patients.1 A similar experience was reported by Swenson et al when they sampled 1.359 medical patients who had been asked by an internist to take the MMPI.-'They found that 89.2% completed the test with no difficulty. 5.8% either refused to cooperate or did not complete the test, and 5% were unable to cooperate (nonreaders. acutely ill, etc.). They concluded that patients accept the idea of completing a personality inventory as part of a medical evaluation. They also surveyed 158 physicians who had used the MMPl with all of their patients. This 14-item questionnaire produced a 70% to 85% endorsement of the test utility. The MMPl therefore appears to be accepted by both physicians and patients as part of a medical evaluation.

In the early literature, several investigators'-4 presented case studies that illustrated the use of the MMPI with medical patients but Guthrie5 was the first to study the use of the test in a medical setting systematically. After an internist had taken the history and recorded his impression, Guthrie classified the MMPI profiles according to two-point codes (a listing of the highest two clinical scales). He studied the medical records of 365 men and 739 women in an attempt to develop a summary of the most frequently appearing physical and psychologic symptoms. Summary descriptions for five MMPI code types were developed and then evaluated by using 66 patients from the practice of a different internist. The findings supported the original descriptions.

Because previous studies of the MMPI in medical patients had used either a small number of subjects or a highly selected sample, psychologists at the Mayo Clinic were reluctant to apply these results to their patient group. Using automated techniques they collected a large number of MMPIs from medical patients; for a 3-year period all patients seen by a group of internists at the Mayo Clinic were given the MMPi.6 The subjects were ambulatory outpatients. 16 years of age or older, who had sufficient reading ability to complete the MMPI. The project continued until valid MMPI profiles of 50,000 medical patients had been collected. It was found that the medical patients had higher scores than the MMPI standardization group on a number of scales. Scores on scales J (Hs), 2 (D), and 3 (Hy). which reflect somatic concern, depressed mood, and denial of emotional problems, were quite elevated relative to the standardization group. Therefore, they concluded that elevations on these scales should be interpreted with caution in a medical population. They believed that these elevations were not necessarily the results of emotional or physical illness. Swenson found that the mean profile of 835 males with negative physical examinations also exhibited elevations on scales Hs and Hy.7 He concluded that, when the MMPI is taken as part of a medical evaluation, elevations on scales Hs and Hy of the MMPI can be expected and should be interpreted conservatively.

Colligan and Osborne examined the MMPl profiles of 1,193 adolescent medical patients (part of the original Mayo Clinic sample) and concluded that these medical patients are different from other groups of adolescents in much the same way that adult medical patients are different from the adult population at large.8 As adolescent medical patients get older, their mean MMPI profile approaches the mean profile for the adult medical patient.

IDENTIFYING FUNCTIONAL ILLNESS

The use of the MMPl to distinguish between functional and organic illness has been reported frequently. This may be the most common use of the MMPI in medical practice. The fact that large numbers of patients in a physician's practice have primary psychologic problems or psychologic components of their physical illness has stirred much interest in finding a simple method of identifying these patients.

In an early investigation, Lairand Trapp studied MMPI profiles of patients in three groups: 20 patients with physical illness, 20 with psychophysiologic complaints, and 20 with neurotic problems.9 They matched the groups for age, education, and intelligence. All three groups had similar MMPI patterns. They found that neurotic patients obtained the highest scores on scales Hs, D. and Hy. The patients with psychophysiologic disorders obtained the next highest scores, and the physically ill patients had the lowest scores. Even the scores of this last group, however, were significantly elevated above normal levels. However, although these groups differed in MMPI scores, the large amount of overlap between the groups made it difficult to use the MMPI to predict the group to which an individual patient belonged.

Gilberstadt and |ancis pursued this question further by attempting to differentiate between organic and functional groups by use of the 13/31 MMPI profile - highest scores on scales 1 (Hs) and 3 (Hy).10 This profile type, sometimes called the conversion V, was thought to indicate the presence of functional complaints. They found a high frequency of both psychologic and physical illness in their 13/31 group. They concluded that it was not feasible to distinguish between patients with functional and organic complaints by using this code type.

Schwartz and Krupp evaluated the possibility that among patients with a 1 3/3 1 code type those with higher scores might be more likely to exhibit functional complaints." From the 50.000 MMPIs collected at the Mayo Clinic they selected profiles of the 13/31 type from 60 male and 60 female subjects. They divided these profiles into three groups on the basis of the scores on scales 1 (Hs) and 3 (Hy): high, medium, and low groups, each with 20 males and 20 females. Examination of the medical records indicated that there were no significant differences between the three groups in the number of patients who received functional diagnoses.

Schwartz et al suggested that consideration of the age and sex of the patient might improve the prediction of medical diagnosis from the 13/31 type.12 In a randomly selected sample consisting of 178 medical patients who had 13/31 type profiles, they found that age and sex did affect the ability of the profile to predict functional or organic diagnosis. Younger men with a 13/31 profile had a fairly even chance of a psychologic, mixed, or organicdiagnosis. Older men with the 13/31 profile were highly likely to fall into the organic category. Older women had a fairly even chance of an organic, mixed, or psychologic diagnosis, but younger women with a 13/31 profile were much more likely to have a psychologic diagnosis. The investigators concluded that a prediction of functional or organic diagnosis cannot be made for either a younger male or an older female with a 13/31 profile. A younger (age 45 or less) female with this profile could be predicted to have a psychologic diagnosis. An older male with this profile could be predicted to have an organic diagnosis.

Calsyn et al, working with patients who had low back symptoms, noted that their patients' complaints were neither all functional nor all organic. I3 Their largest group of patients had organic findings that were clearly established but not sufficient to explain the reported degree of pain behavior. They suggested that the tendency to dichotomize between functional and organic pain should be de-emphasized. They indicated that both organic and psychologic factors in pain patients combine to produce the observed level of pain behavior. They formed two groups: an organic group in which the organic cause was judged to be appropriate to the pain level, and a mixed group in which the organic basis for pain did not account fully for the pain level. They found that the mixed group had higher scores on scales 1 ( Hs) . 2 (D), and 3 (Hy).

Recently, several authors have attempted to use the MMPI to separate organic from functional erectile dysfunction. Beutler et al studied 15 men who had psychogenic erectile dysfunction and 17 who had organic erectile dysfunction.14 From inspection of the MMPI profiles they found that profiles with a T score greater than 60 on scale 5(M0 and any T score greater than 70 on the MMPI indicated psychogenic impotence. They reported that use of these rules correctly classified 90% of their sample. Other authors have been unable to replicate this finding.15,16 A typical finding was that by Osborne and Furlow who reported that use of this index led to 39% of the sample of 100 impotent men being classified incorrectly.17

In summary, many studies have found that patients with functional diagnoses obtain higher scores on scales 1 (Hs), 2 (D), and 3 (Hy) compared to patients with organic diagnoses, but these differences cannot be relied upon to identify individual members of either group. Most of the attempts to use the MMPl to make this distinction have resulted in too large a number of misclassifications. Consideration of age and sex along with the MMPI profile can help in this discrimination. It is also possible that specific items in the MMPI might aid in the identification of organic versus functional diagnoses. As Gilberstadt and fancis suggested, this distinction in part might be a function of factors to which personality tests are not particularly sensitive.10

RESPONSE TO TREATMENT

The assumption that personality variables could influence a patient's compliance with and response to medical and surgical treatment has led a number of investigators to use the MMPI to identify these variables. Maruta et al attempted to predict which patients might benefit from a behaviorally oriented rehabilitation program for chronic pain.18 They examined differences between groups of patients who were judged, at 1-year follow-up. to have succeeded or failed in a pain management program. The statistical evaluation indicated that many non-test factors predicted success or failure. Duration of pain, work time lost, number of surgical procedures, subjective pain level, and drug dependency all differentiated between the two groups. MMPI T scores greater than 80 on scales 1 (Hs) and 3 (Hy) also were associated with failure in the program.

Kutner studied the effect of personality factors on recovery from cholecystectomy.19 His criterion was the number of postoperative doses of analgesie medication administered. At the time his data were collected, surgeons generally were ordering postoperative medication on a "need" basis. Kutner considered using several other criteria but could not use them for various reasons. Some studies have examined the number of days between operation and dismissal from the hospital. Unfortunately, dismissal often is an administrative rather than a medical decision. Postoperative time until ambulation is not a useful variable because patients arc strongly encouraged to walk as soon as possible after most types of operation. There is little variation when a given type of operation is considered.

The MMPI clinical scales did not enable Kutner to discriminate between patients who received large and small numbers of doses of analgesics. He found, however, that groups of MMPI items were able to discriminate between the two groups. A factor analysis of these items yielded the following results. Patients who obtained high scores on a factor that appeared to be measuring sociopathic tendencies requested large numbers of doses. Those who obtained high scores on a factor measuring "worry" were also likely to require large numbers of doses. High scores on the factor "moralistic thinking" used fewer doses. It appeared, therefore, that conventional, moralistic, and nonworrying patients required less analgesic medication after operation than those who scored on the opposite end of these dimensions.

EMOTIONAL ASPECTS OF ILLNESS

In their study of the relationship between repressionsensitization and physical illness, Byrne el al hypothesized that both sensitizers and repressors would have more frequent and more severe illnesses than people occupying the neutral position between those two groups.20 The repression-sensitization scale from the MMPI and a health survey were administered to two samples of college students. The hypothesis was not supported. Sensitizers had a greater frequency and severity of psychosomatic complaints than repressors. These findings suggested that repressors and sensitizers either have physiologic differences affecting susceptibility to illness or that they have different ways of viewing and dealing with an illness. A sensitizer, for example, might be more likely to go to a clinic with a physical complaint than a repressor who might avoid dealing with the complaint.

In examining the relationship between MMPI scores and disability in patients with multiple sclerosis. Davis et al found that significant disability was more frequent among patients whose Hy score on the MMPI exceeded their Pt score than among those whose Pt score was higher than their Hy score.21 Because that was a correlation study, it is not known whether this difference in MMPI scores was the cause or the result of degree of disability.

Dirks et al studied the panic-fear dimension of the Asthma Symptom Checklist.22 A MMPl scale, the panicfear scale, was developed on the basis of this dimension and has been used to investigate intractibility and chronicity in respiratory disease. They found that patients scoring high on the MMPI panic-fear scale were more often assigned larger and more frequent doses of medication at dismissal than were those with low scores even though mean levels of pulmonary function did not differ. The panic-fear scale was also found to have a relationship to length of hospitalization in patients with respiratory difficulties. 23

MEAN PROFILES OF PATIENT GROUPS

Many investigations of the MMPI in medical patients present the mean MMPI profile of patients with a given disease. These profiles are often compared with those of a normal control group or a medical control group. Generally, although these articles identify significant differences between the diagnostic group and the control group, there usually is a high degree of overlap between the groups. This makes it difficult to use the information for clinical work with an individual. It has also become apparent that there is a high degree of similarity in the MMPI profiles of patients with various diseases. In most of these studies the MMPl was given to patients who already had a disease. Therefore, we do not know if the scores reflect personality factors that contribute to the development of the disease or factors that are reactions to the illness.

Diseases for which mean profiles have been presented include ulcerative colitis,24 rheumatoid arthritis,25,26 myasthenia gravis,27 multiple sclerosis,2830 Huntington's chorea.51 menstrual irregularity,52 female urethral syndrome,35 and chronic pain.54

In most groups of medical patients, scales 1 (Hs) and 3 (Hy) are elevated; in many groups, scale 2 (D) is also high in comparison to control groups. These scales appear to be highest among patients with chronic disease and somewhat lower among general medical patients. Even patients in a medical setting who have come for a routine medical examination and have no physical complaints produce slightly elevated scores on these scales.

CONCLUSION

McKinley and Hathaway originally had hoped that the MMPl would help distinguish between functional and organic complaints. Unfortunately, the large overlap between patient groups and control groups has made it difficult to use the MMPI in making this distinction for a specific patient. The patient who, in response to emotional stress, uses defense mechanisms that lead to functional illness uses these same mechanisms to cope with the stress of physical illness. Therefore, even though the MMPI can identify the personality style, it has no way of determining whether or not physical illness is present.

A more profitable use of the MMPI has been as a screening device to identify medical patients who have emotional difficulties. The test has been used extensively for this purpose in clinical settings, and it is a task for which the MMPI is well suited. The identification of maladaptive defense mechanisms and effective coping strategies can lead to more effective patient care.

REFERENCES

1. McKinley IC. Hathaway SR: The identification and measurement of the psycho neuroses in medical practice. The Minnesota Multiphasic Personality Inventory JAMA 1945; 122:101-167.

2. Swenson WM. Rome HI'. Pearson IS. et al: A totally automated psychological test: Experience in a medical center. IAMA 1965: 191:925-927

3. Houk TW: Use of the MMCl in diagnosis of psychoneuroses. Northwest Medicine 1946: 45:248-252

4 Kamman OR: Psychosomatic diagnosis Lancet 1947: 67: 102-107.

5 Guthrie GM A Study of the Personality Characteristics Associated With Ditsor den, Encountered by an Internist, thesis. University of Minnesota. Minneapolis, 1950.

6 Swenson WM. Pearson IS. Osborne I): An MMPI Source Book Basic Item Scale, and Pattern ilata on 50,000 Medical Patients Minneapolis. University of Mm nesota Press. 1975.

7. Swenson WM: Automated personality assessment in medical practice. Med Clin North Am 1970; 54:855-849.

8 Colligan RC. Osborne D: MMPI profile- from adolescent medical patients. I Clin Psychol 1977: 55:186-189

9. Lair CV, Trapp F.P: The differential diagnostic value of MMPI with somatically disturbed patients. J CVm Psychol 1962; 18: 146-147

10, Gilberstadl H. Jancis M: "Organic" vs "functional" diagnoses (rom 1-5 MMPI profiles. J Clin Psychol 1967; 25:480 485

11 Schwartz MS. Krupp NK: The MMPI "conversion V" among 50.000 medical patients: A study of incidence, criteria, and profile elevation. J < lin Psychol 1471. 27:89-95

12 Schwartz MS. Osborne D. Krupp NI Moderating ellects ol age and sex on lhe association of medical diagnoses and 1-5 51 VlMPI profiles. J Clin Pschol 1972; 28502-505.

13 Calsyn DA. l.ouks I. Freeman CW . The use ot the MMPl with chronic low back pain patients with a mixed diagnosis. J Clin Psychol 1976: 52:552-556

14 Beutler I I., Karacan I. Auch VM. el al: MMPI and MIT discriminators of biogenic and psychogenic impotence J Consult Clin Psychol 1975; 45:899-905

15 Marshall P. Surridge D. Delva N Differentiation ol organic and psychogenic impotence on lhe basis of MMPI decision rules. J Consult Clin Psychol 1980; 48:407-408

16 Robiner WN. Codec Cl, Cass AS. et al: The role ol Minnesota Multiphasic Personality Inventory in evaluation of erectile dysfunction. J Urol 1982: 128:487-488.

17 Osborne D, Furlow WI.: Use of the MMPI Io identity psychogenic impotence Presented at the American Urological Association, las Vegas. April 17-21. 1985.

18. Maruta T, Swanson DW Swenson WM: Chronic pain: Which patients mas a pain-managemenl program help'1 Pain 1979; 7 521-529.

19 Kutncr I A Psychological factors Influencing Pain and Recoven I rom Surgery thesis. University ol Minnesota. Minneapolis. 1478.

20. Byrne D. Steinberg MA. Schwartz MS: Relationship between repression-sensitization and physical illness. J Abnorm Psychol 1968; 75: 1 54- 1 55.

21 Davis LJ Jr. Osborne D. Siemens PI. et al. MVIPI correlates with disability in multiple sclerosis. Psychol Rep 1971; 28:700-702

22. Dirks JF, Jones NF, Kinsman A. Panic-tear: A personality dimension related to intractability in asthma Psychosom Med 1977; 59:120-126

25 Dirks JF. Kinsman RA. Jones NF. et al. Panic (ear: A personality dimension related Io length of hospitalization in respiratory illness J Asthma 1477. 14:61-71

24. West Kl.: MMCI correlates of ulcerative colitis. J Clin Psychol 1470; 26:214 219

25. Naiven FB. O'Brien JF: Personality patterns of rheumatoid arthritis patients Arthritis Rheum 1964; 7:18-28.

26. Colley HF. Swenson WM. Slemhilber RM: Personality characteristics of patients with rheumatoid arthritis- Psychosomatic! 1970; I 1.45-49

27. Schwartz ML. C'ahill R: Personality assessment in myasthenia gravis with the MMPI Percept Mol Skills 1970:51:766.

28. Canter AH: MMPI profiles in multiple sclerosis. J Consult Clin Psychol 1951; 15:255-256

29 Shontz FC MVlPI responses of patients with multiple sclerosis J Consult Clin Psychol 1955; 19:74.

50 Matthews CG. Cleeland CS. Hopper CL: Neuropsychological patterns in multiple sclerosis. Diseases of the Nenvus System 1970; 51:161-170.

51. Boll DW, Heaton R. Rcilan RM: Neuropsychological and emotional correlates ol Huntington's chorea. J Nerv Ment Dis 1974; I 58:61-69.

52. Hain ID. Linton PH. I her HVV. et al Menstrual irregularity, symptoms and personalis I Psychosom Res 1970: 14:81-87

33. Carson CC III. Osborne D. Segura JW: Psychologic characteristics of patients with female urethral syndrome. j Clin Psychol 1979; 55:512 51 5

54 Swanson DW. Swenson WM. Manila I. et al: Program lot managing chronic pain. I Program description and characteristics ol patterns Mayo Clin Proc 1976; 51 401-408

10.3928/0048-5713-19850901-06

Sign up to receive

Journal E-contents