it was this kind of research which helped both preserve and extend the National Health Insurance benefit for the mentally ill in West Germany and led to a more efficient and effective prior authorization and peer review system throughout that country.

The fourth study is a more recent study by the Research Department of Blue Cross and Western Pennsylvania which assessed the medical/surgical utilization of a group of subscribers (N:169) who used a psychotherapy outpatient benefit in community mental health centers with a comparison group of subscribers. The findings showed that medical/surgical utilization was reduced significantly for the group who used the psychiatric benefit. The monthly cost per patient for medical services dropped by $9.41, from $16.47 to $7.06. This was well below the average per capita cost for the control group. Both medical/ surgical inpatient days per month and outpatient visits per month were down by more than 54 percent. Further, this phenomenon or reduced medical/surgical utilization with exposure to outpatient psychotherapy was found to be independent of age, sex, or employment level.

The above four studies suggest that significant benefits accrue to an insurance or prepaid system when mental health benefits are available and accessible. Any discussion about coverage for mental health under national health insurance must take these findings into account.

STEVEN S. SHARFSTEIN, M.D.

[From Medical Care, January-February 1967, Vol. V, No. 11

PSYCHIATRIC SERVICES AND MEDICAL UTILIZATION IN A
PREPAID HEALTH PLAN SETTING

(By William Follette, M.D.,* and Nicholas A. Cumming, Ph.D.**) In two previous studies [5, 6] the psychiatric practitioner's contention that emotionally disturbed patients do not seek organic treatment for their complaints following the intervention of psychotherapy have been investigated. Although it has long been recognized that a large number of the physical complaints seen by the physician are emotionally, rather than organically, determined, the more precise relationship between problems in living and their possible expression through apparent physical symptomatology has been difficult to test experimentally. As noted in the previous study, the GHI Project [1] demonstrated that users of psychiatric services were also significantly frequent users of medical services, but the Project was not able to answer the question of whether there is a reduction in the use of medical services following psychotherapy.

Because the facilities and structure of the Kaiser Foundation Health Plan accord an experimental milieu not available to Avnet, the original pilot project in San Francisco was able to demonstrate a significant reduction in medical utilization between the year prior to psychotherapy, and the two years following its intervention. Certain methodologic problems inherent to the pilot study indicated caution and the need for refinement and replication to avoid arriving at premature conclusions. The lack of a control group of what might be termed psychologically-disturbed high-utilizers who did not receive psychotherapy was a serious omission in the first experiment. † Furthermore, an error in the tabulation of in

5 Jameson, John, Shuman, Larry J., and Young. Wanda W. "The effects of Outpatient Psychiatric Utilization on the Costs of Providing Third-Party Coverage. Res. Series 18. December. 1976.

Chief Psychiatrist, Kaiser Foundation Hospital and the Permanente Medical Group, San Francisco, Calif.

**Chief Psychologist, Kaiser Foundation Hospital and the Permanente Medical Group, San Francisco, Calif.

Presented at one of the Contributed Papers Sessions sponsored by the Medical Care Section at the 94th Annual Meeting of the American Public Health Association, San Francisco, Calif., Oct. 31-Nov. 4, 1968.

This study was primarily financed by Grant PH 108-64-100 (P), U.S. Public Health Service. The authors gratefully acknowledge the assistance and cooperation of Mr. Royal Crystal. Deputy Chief. Health Economics Branch. Secondary financial support for this study was through Grant No. 131-7241. Kaiser Foundation Research Institute.

This paper is a report of the first of two investigations seeking to develop and test methods of assessing the effect of psychiatric services on medical utilization in a comprehensive medical program. Part II deals with prospective, rather than respective, methodology, and will be reported later.

The authors acknowledge their debt to Dr. M. F. Collen for this and other suggestions, and to Mr. Arthur Weissman. Medical Economist, Kaiser Foundation Medical care entities, for his expert consultation.

patient utilization was discovered after the experiment had been concluded. †† In addition, the question was raised whether the patients studied might, subsquent to the two years following psychotherapy, revert to previous patterns of somatization or, as a new pattern, merely substitute protracted and costly psychotherapy for previous medical treatment.

THE PROBLEM

This study investigated the question of whether there is a change in patients' utilization of outpatient and inpatient medical facilities after psychotherapy, comparing the patients studied to a matched group who did not receive psychotherapy.

Psychotherapy was defined as any contact with the Department of Psychiatry, even if the patient was seen for an initial interview only. The year prior to the initial contact was compared with the five subsequent years in both groups. The problem can be stated simply: Is the provision of psychiatric services associated with a reduction of medical services utilization (defined as visits to other medical clinics, outpatient laboratory and x-ray procedures, and days of hospitalization)?

METHODOLOGY

The setting.-The Kaiser Foundation Health Plan in the Northern California Region is a group-practice prepayment plan offering comprehensive hospital and professional services on a direct service basis. Professional services are provided by the Permanente Medical Group-a partnership of physicians. The Medical Group has a contract to provide comprehensive medical care to the subscribers, of whom there were more than a half million at the time of this study. The composition of the Health Plan subscribers is diverse, encompassing most socioeconomic groups. The Permanente Medical Group comprises all major medical specialties; referral from one specialty clinic to another is facilitated by the organizational features of group practice, geographical proximity and use of common medical records. During the years of this study (1950-1964), psychiatry was essentially not covered by the Northern California Health Plan on a prepaid basis, but in some areas of the Northern California region psychiatric services were available to Health Plan Subscribers at reduced rates. During the six years of the study, the psychiatric clinic staff in San Francisco consisted of psychiatrists, clinical psychologists, psychiatric social workers, resident psychiatrists at the third- or fourth-year level, and psychology interns, all full-time. The clinic operates primarily as an out-patient service for adults (age eighteen or older), for the evaluation and treatment of emotional disorders, but it also provides consultation for non-psychiatric physicians and consultation in the general hospital and the emergency room. There is no formal "intake" procedure, the first visit with any staff member being considered potentially therapeutic as well as evaluative and dispositional. Regardless of professional discipline, the person who sees the patient initially becomes that patient's therapist unless there is a reason for transfer to some other staff member, and he continues to see the patient for the duration of the therapy. An attempt is made to schedule the first interview as soon as possible after the patient calls for an appointment. There is also a "drop-in" or non-appointment service for emergencies so that patients in urgent need of psychiatric help usually can be seen immediately or at least within an hour or two of arrival at the clinic.

One of the unique aspects of this kind of associated health plan and medical group is that it tends to put a premium on health rather than on illness, i.e., it makes preventive medicine economically rewarding, thereby stimulating a constant search for the most effective and specific methods of treatment. The question of how psychiatry fits into comprehensive prepaid medical care is largely unexplored; there are not many settings in which it can be answered. Another feature of group practice in this setting is that all medical records for each patient are retained within the organization.

Subjects. The experimental subjects for this investigation were selected systematically by including every fifth psychiatric patient whose initial interview took place between January 1 and December 31, 1960. Of the 152 patients thus selected, 80 were seen for one interview only, 41 were seen for two to eight

At that time days of hospitalization per patient and by year were tabulated from each patient's outpatient medical records. Subsequent investigation has revealed that only about a third of the outpatient charts reviewed contained summaries of hospital admissions, and that tabulation of inpatient utilization must be made directly through the separatelykept inpatient records.

interviews (mean of 6.2) and were defined as "brief therapy," and 31 were seen for nine or more interviews (mean of 33.9) and were defined as "long-term therapy."

Thus, each experimental patient was matched with a control patient in the criteria above, but without reference to any other variable. Both samples ranged in age from 24 to 62, with a mean of 38.1. Of these, 52 percent were women and 63 percent were blue-collar workers or their dependents. The satisfaction of so many criteria in choosing a matched control group proved to be a tedious and time-consuming procedure.

Review of the medical records of the psychiatric sample disclosed consistent and conceptually useful notations in the year prior to the patients' coming to psychotherapy, which could be considered as criteria of psychological distress. These consisted of recordings, made by the physicians on the dates of the patients' visits, which were indicative of those patients' emotional distress, whether or not the physicians recognized this when they made the notations. These (38) criteria were assigned weights from one to three in accordance with the frequency of their appearance in medical records and in accordance with clinical experience about the significance of the criteria when encountered in psychotherapeutic practice. The criteria, with weights assigned, are presented in Table 1. In com paring the charts of the psychiatric patients with those of Health Plan patients randomly drawn, it was determined that although some criteria were occasionally present in the medical records of the latter, a weighted score of three within one year clearly differentiated the psychiatric from the non-psychiatric groups. Accordingly, therefore, in matching the control (non-psychotherapy) group to the experimental (psychotherapy) group, the patients selected had records which indicated scores of three or more points for the year 1959. The mean weights of the three experimental groups and the control group in terms of the 38 criteria of psychological distress are presented in Table 2: note that there was no significant difference between this dimension of the two groups in 1959.

TABLE 1-CRITERIA OF PSYCHOLOGICAL DISTRESS WITH ASSIGNED WEIGHTS

1 point

2 points

3 points

1. Tranquilizer or sedative requested. 23. Fear of cancer, brain tumor, vene

34. Unsubstantiated complaint there is something wrong with genitals. 35. Psychiatric referral made or requested.

36. Suicidal attempt, threat, or preoccupation.

37. Fear of homosexuals or of homosexuality.

38. Nonorganic delusions and/or haltucinations; paranoid ideation; psychotic thinking or psychotic behavior.

1 Refers to the last 4 questions (relating to emotional distress) on a Modified Cornell Medical Index-a general medical questionaire given to patients undergoing the multiphasic health check in the years concerned (1959–62).

TABLE 2.-SCORES FOR CRITERIA OF PSYCHOLOGICAL DISTRESS, FOR THE EXPERIMENTAL GROUPS AND THE CONTROL GROUP DURING THE YEAR PRIOR TO PSYCHOTHERAPY (1959)

In order to facilitate comparison of the experimental (psychotherapy) and control (non-psychotherapy) groups, one last criterion for inclusion in the matched group was employed. Each subject in the control group had to be a Health Plan member for the first three consecutive years under investigation inasmuch as the experimental group, though demonstrating attrition in continued membership after that time, remained intact for those years.

Dependent variable.-Each psychiatric patient's utilization of health facilities was investigated first for the full year preceding the day of his initial interview, then for each of the succeeding five years beginning with the day after his initial interview.

The corresponding years were investigated for the control group which, of course, was not seen in the Department of Psychiatry. This investigation consisted of a straightforward tabulation of each contact with any outpatient facility, each laboratory report and x-ray report. In addition a tabulation of number of days of hospitalization was made without regard to the type or quantity of service provided. Each patient's utilization scores consisted of the total number of separate outpatient and inpatient tabulations.

RESULTS

The results of this study are summarized in Table 3. which shows the differences by group in utilization of outpatient medical facilities in the year before and the five years after the initial interview for the psychiatric sample, and the utilization of outpatient medical services for the corresponding six years for the non-psychotherapy sample.

TABLE 3.-UTILIZATION OF OUTPATIENT MEDICAL SERVICES (EXCLUDING PSYCHIATRY) BY PSYCHOTHERAPY GROUPS FOR THE YEAR BEFORE (1-B) AND THE 5 YR AFTER (1-A, 2-A, 3-A, 4-A, 5-A) THE INITIAL INTERVIEW, AND THE CORRESPONDING YEARS FOR THE NONPSYCHIATRIC GROUP

The data of Table 3 are summarized as percentage in Table 4, which indicates a decline in outpatient medical (not including psychiatric) utilization for all three psychotherapy groups for the years following the initial interview, while there is a tendency for the non-psychotherapy patients to increase medical utilization during the corresponding years, Applying t-tests of the significance of the standard error of the difference between the means of the "year before" and the means of each of the five "years after" (as compared to the year before), the following results obtain. The declines in outpatient (non-psychiatric) utilization for the "one session only” and the "long-term therapy" groups are not significant

for the first year following the initial interview while the declines are significant at either the .05 or .01 levels for the remaining four years. In the "brief therapy" group, there are statistically significant declines in all five of the years following the initial interview. As further indicated in Table 4, there is a tendency for the control group to increase its utilization of medical services, but this proved significant for the "fourth year after" only.

TABLE 4-COMPARISON OF THE YEAR PRIOR TO THE INITIAL INTERVIEW WITH EACH SUCCEEDING YEAR, INDICATING PERCENT DECLINE OR PERCENT INCREASE (LATTER SHOWN IN PARENTHESES) IN OUTPATIENT MEDICAL (NONPSYCHIATRIC) UTILIZATION BY PSYCHOTHERAPY GROUPING, AND CORRESPONDING COMPARISONS FOR THE CONTROL GROUP, WITH LEVELS OF SIGNIFICANCE

The question was raised as to whether the patients demonstrating declines in medical utilization have done so because they have merely substituted protracted psychotherapy visits for their previous medical visits.

As shown in Table 5, the number of patients in the one-session-only group who return in the third to fifth years for additional visits is negligible. Comparable results are seen in the brief-therapy group. In contrast, the long-term-therapy group reduces its psychiatric utilization by more than half in the "second year after," but maintains this level in the succeeding three years. By adding the outpatient medical visits to the psychiatric visits, it becomes clear that whereas the first two psychotherapy groups have not substituted psychotherapy for medical visits, this does seem to be the case in the long-term psychotherapy group. These results are shown in Table 6, and indicate that the combined outpatient utilization remains about the same from the "year before" to the "fifth year after" for the third psychotherapy group, while declines are evident for the first two psychotherapy groups. As regards the combined (medical plus psychiatric) utilization, the long-term psychotherapy group is not appreciably different from the control (non-psychiatric) group.

TABLE 5.-AVERAGE NUMBER OF PSYCHOTHERAPY SESSIONS PER YEAR FOR 5 YEARS BY EXPERIMENTAL GROUP

TABLE 6.-COMBINED AVERAGES (OUTPATIENT MEDICAL PLUS PSYCHOTHERAPY VISITS) OF UTILIZATION BY YEARS BEFORE AND AFTER PSYCHOTHERAPY FOR THE EXPERIMENTAL GROUPS, AND TOTAL OUTPATIENT UTILIZATION BY CORRESPONDING YEARS FOR THE CONTROL (NONPSYCHIATRIC) GROUP