SETTING AND NATURE OF SHORT-TERM MENTAL HEALTH BENEFIT

The Group Health Association of Washington, D.C. is a comprehensive prepaid group practice program whose participating population resides almost entirely in the metropolitan Washington, D.C. area and is comprised of three groups: federal government employees, D.C. transit workers, and general members. In November 1964, GHA included a limited outpatient mental health benefit in its structure of benefits for its government employee group who then comprised 66 percent of the GHA participant population of approximately 54,000. In January 1965, this benefit was extended to the general members who accounted for 18 percent of the participant population. Thus, 84 percent of the GHA population had some coverage for short-term outpatient psychiatric care at the inception of the partially prepaid benefit. Acute short-term hospital care had previously been part of the benefit structure.

At its initiation, the GHA mental health benefit offered under prepayment was essentially as follows: GHA paid up to 15 dollars for each of 10 therapy sessions in a membership-year for outpatient treatment of acute mental illness and emotional disorders subject to significant improvement through short-term outpatient therapy.1 A GHA screening psychiatrist determined eligibility for referral on benefits. When the patient was referred by a GHA nonpsychiatric physician to the GHA screening psychiatrist for evaluation purposes as to eligibility for benefits, there was no charge to the patient for that visit or visits. During the study period, a patient could also self-refer to the screening psychiatrist. An evaluation of the patient's psychiatric condition was made by the screening psychiatrist and, on the basis of his diagnostic impression, he recommended appropriate psychiatric care where indicated, and he determined whether GHA coverage for benefits could be approved. If short-term therapy was authorized under the benefit, the patient was referred to psychiatrists or other mental health disciplines. If the condition was chronic, and hence not covered by the benefit, referral could still be made to another agency or psychiatrist, but no payment would be made by GHA for such care.

STUDY DESIGN

The basic study plan was to compare, for the case group under study, the utilization of GHA medical services before and after each patient was referred on benefits for short-term outpatient psychiatric therapy. The "before" period was the 12-month interval immediately preceding the date of referral by the screening psychiatrist. It was considered likely that virtually all of the patients undergoing therapy would have completed such care during the first three months immediately following referral. Since such therapy was apt to affect the utilization of GHA services during this period, and to allow sufficient time for completion of the therapy on benefits, the "after" period of 12 months' duration was taken to begin three months following the referral date. Thus, the records for each study patient were reviewed for a 27-month exposure period, although the three-month "psychiatric therapy interval" was not to be included in the "beforeafter" analysis of medical services utilization.

For purposes of the study, it was desirable that the study group be confined to persons who were covered by the same mental health benefit. It was, therefore, decided to limit the study group to all patients enrolled under the "high option" or "premium" plan who were referred on benefits for psychiatric therapy during the first year the benefit structure was in effect. Thus, as a by-product, the results of the study could provide a baseline for any future studies based on a revised benefit structure. (After the first year, the GHA mental health benefit substantially increased and broadened.) [7]

1 The limit of 10 therapy sessions was a renewable benefit each membership-year (i.e., year beginning with each anniversary date of joining the plan). Thus, if therapy was initiated towards the end of one membership-year and carried into the next, the patient could actually have as many as 19 sessions for the same referral.

2 Under the "low option" or "standard" plan, GHA paid up to 10 dollars (as compared with 15 dollars under "high options") per therapy session. Only about 10 percent of the GHA members are enrolled in the "low option" plan.

92-202-77-28

Since the GHA mental health benefit during the study period applied only to the federal employee and general member groups (including covered family members), they comprised the study population. The GHA medical records for these enrollees were reviewed by GHA staff to identify all patients who were referred to, and seen by, the screening psychiatrist during the period November 1, 1964 through October 31, 1965, the first full year in which the psychiatric benefit was-in effect. To protect the confidentiality of the patient, individuals were not identified by name to the study staff. Also, it should be noted that the confidential psychiatric notes are not part of the medical record and were not made accessible for this study.

A total of 726 patients (excluding GHA staff and dependents) were referred to the screening psychiatrist. Of this total, 409 patients were excluded from the study because they were judged ineligible for coverage under this benefit or because they overtly refused psychiatric care. Specifically, 161 were judged not to be in need and hence not referred for outpatient psychiatric care; 197 were referred for psychiatric care but not on benefits; referral was deferred for 45 patients; and 6 patients who would have been referred on benefits overtly refused to accept such care.

The records for the remaining 317 patients seen by the screening psychiatrist were reviewed for the 27-month period referred to earlier. From this total, 61 were eliminated from the study as follows: 57 cases were not available for the full 27-month period (35 began membership less than one year prior to the date seen by the screening psychiatrist, and 22 terminated their membership within the 15-month period following that date); for four patients the files were not available. This left 256 patients who comprised the study group. Of the final study group, 197 were enrolled in the federal employee program and 59 were general members-approximately in the same ratio to one another that these two groups comprised in the total GHA population.

As point of interest, the age distributions were examined for the 409 patients ineligible for benefits and the 61 eligibles who did not otherwise meet the study criteria. The age distribution for the former group was found to be very similar to that of the 256 study patients; however, the latter group of 61 patients had a somewhat younger age distribution than the final study group.

Aata extracted from the medical records were counts of all visits to GHA physicians for medical care, all visits for x-ray and laboratory procedures, as well as the number of visits made for psychiatric therapy under the mental health benefit. Other data abstracted for each patient, where available, were age, race, sex, and psychiatric diagnostic impression. Information on psychiatric and nonpsychiatric hospitalizations recorded in the medical record was also extracted. However, study data on hospitalizations were incomplete because such information was not generally recorded on patients who were hospitalized outside of GHA auspices. Also, during the period of study, the GHA hospitalization information was not consistently available in the progress notes which formed the primary source of data for this pilot study.

RESULTS

The distribution of the study population by age, sex, and race is shown in Table 1. Approximately 70 percent of the study group were from 25 to 64 years of age at time of referral on psychiatric benefits. In contrast, only 50 percent of the total GHA participant population (in the federal employee and general groups) were in this age group during the study period. [3] About 60 percent of the study group were female, which was slightly higher than the proportion of females in the total GHA membership. With respect to race, about 83 percent of the study group were Caucasian. Although no precise data on race are available for the total GHA membership, the proportion of Caucasians in the total membership is estimated to have been appreciably less than that in the study group. Specific psychiatric diagnosis for each patient was not uniformly recorded in the medical records. However, from information which was re

corded, based on the evaluation of the screening psychiatrist or the psychiatrist providing therapy it was possible to classify the psychiatric diagnostic impression into broad categories for three fourths of the study group. Among those for whom the diagnostic impression was determined, 21 per cent were classified psychotic, 55 per cent psychoneurotic, 11 per cent with personality disorders, 11 per cent as having a transient situational personalty disorder, and 2 percent were considered to have some other psychiatric problem.

TABLE 1.-DISTRIBUTION OF STUDY GROUP BY AGE, RACE, SEX, AND PSYCHIATRIC DIAGNOSTIC IMPRESSION

Initially, the data were analyzed separately according to the specific medical department or ancillary service in which the patients were seen (i.e., internal medicine, other nonpsychiatric medical department, laboratory, x-ray). Almost 95 per cent of the visits by the study patients for physician services were made to the department of internal medicine. However, since the study findings for visits to internal medicine were similar to those for other non-psychiatric medical departments, the data for all medical departments were combined in the analysis presented here. Similarly, with respect to ancillary services, the findings on visits for laboratory procedures were essentially the same as those for x-ray visits, so the data for laboratory and x-ray services were also combined.

Study findings presented below compare separately the physician and ancillary (laboratory or x-ray) services received by the study group during the 12-month periods before and after referral on psychiatric benefits, by age, race, sex, diagnosis, and number of psychiatric therapy sessions attended on benefits. It was not possible to conduct a "before-after" analysis with respect to utilization of psychiatric services. Although some psychiatric counseling was provided on a feefor-service basis prior to the initiation of the mental health benefit, there was no psychiatry department as such at GHA at that time and, therefore, no comparable or meaningful basis for comparison. Thus, the "before-after" analysis was limited to utilization of nonpsychiatric medical services.

Table 2 shows the number of study patients who received care from the various GHA departments, except psychiatry, and the number of visits made to these departments during the "before" and "after" periods. Also shown is the per cent decrease from the "before" to the "after" period with respect to number of patients seen and number of visits made. Each visit for laboratory or x-ray services was counted only once regardless of the number of procedures performed at each visit.

TABLE 2.-COMPARISON OF NUMBER OF PATIENTS SEEN AND VISITS MADE DURING YEAR BEFORE AND YEAR AFTER PSYCHIATRIC REFERRAL BY TYPE OF SERVICE

1 Each visit for laboratory or X-ray services was counted only once regardless of the number of procedures performed at each visit.

It is clearly evident from these data, in terms of persons seen and visits made, that medical and ancillary services were each provided to more of these patients and more frequently before psychiatric referral than after. Thus, the reduction in the number of patients seen by the nonpsychiatric medical departments was 13.6 percent, and for laboratory or x-ray procedures, 15.7 per cent. Similarly, in terms of number of visits made, the reduction was approximately 30 per cent both for physician services and for laboratory or x-ray procedures.

Viewing the reduction in utilization another way, the average (mean) number of visits made by the 256 study patients, during the "before" and "after" periods, respectively, were 4.94 and 3.92 for physician services, and 3.11 and 2.18 for laboratory or x-ray procedures.

3

Overall, the study group experienced a total reduction of some 30 per cent in the number of visits made for physician and ancillary services. The difference between the periods before and after referral with respect to the number of patients seen was statistically significant (P<.001) for physician services as well as for laboratory or x-ray procedures. Similarly, for each of these services, the reduction in the mean number of visits was also statistically significant (P<.001).

The study data were analyzed further to determine whether the observed decreases after psychiatric referral held for various subgroups of the study population. Thus, for both physician services and ancillary services, the "before" and "after" periods were compared with respect to the per cent change in number of persons served and total visits made according to age, race, sex, psychiatric diagnostic impressions and number of psychiatric therapy sessions attended under benefit.

The findings presented in Table 3 clearly show the overall consistency of reduction in utilization of the physician and ancillary services by the study group. Although some variation existed in the extent of decrease (partly due to small numbers, in some cells), the pattern of reduced utilization of these services held throughout each of the distributions. There was particularly little variation in the per cent change by age. It is also of interest to note that patients who did not avail themselves of the short-term outpatient therapy benefit generally showed as great a relative reduction in utilization of medical services as did those who received the full benefit of at least 10 sessions.

3 McNemar's chi-square test for correlated samples was used.

4 The two-tailed t-test of paired (before-after) differences was used.

5 For both of the service categories, statistical tests of significance were performed comparing the various age groups, Caucasians with those of other races, males with females, the various diagnostic categories, and those who had no psychiatric therapy sessions under benefit with those who had 10 or more sessions. With respect to persons seen, each patient was classified as to whether or not he showed a "before-after" reduction in number of visits made, and a chi-square test was used to compare the dichotomous distributions for the various comparison groups. None of these comparisons was significant at the .05 levels. With respect to visits made, either an analysis of variance or a two-tailed t-test was made of the difference between the comparison groups in the mean "before-after" reduction in number of visits. In only one instance (the greater reduction observed among males than females in average number of laboratory or X-ray visits, P<.02) was the observed difference statistically significant at the .05 level.

TABLE 3. PERCENT DECREASE DURING YEAR AFTER REFERRAL AS COMPARED WITH PRIOR YEAR UTILIZATION OF NONPSYCHIATRIC PHYSICIAN SERVICES AND LABORATORY OR X-RAY PROCEDURES, ACCORDING TO PATIENT CHARACTERISTICS AND PSYCHIATRIC THERAPY ON BENEFITS

1 Percent not shown in any cell where base (number before referral) was less than 10.

Another indication of the consistency of reduced utilization of physician and ancillary services after psychiatric referral is evident in the data in Table 4. Here, a determination was made as to whether each patient made fewer, more, or the same number of visits during the 12-month period after psychiatric referral as he or she made during the prior year for physician services or for laboratory or x-ray procedures. Only about one fourth of the study patients made more visits for physician services after referral than before in contrast with the almost 60 percent who made fewer visits after referral. Similarly, only 28 percent of the patients made more visits for laboratory or x-ray procedures after referral than before, while 52 per cent made fewer such visits. Both of these differences were statistically significant (P<.001). When the patients were grouped according to the actual number of visits made in the year preceding referral, this pattern of fewer visits held for virtually all groups of patients who had at least two visits in the prior year for physician or ancillary services. The greatest relative reductions occurred among those who made the most visits during the prior year. Thus, of the 81 patients who made more than five visits for physician services during the year preceding referral 64 (79 per cent) made fewer visits in the post-referral year than they did in the prior year.

TABLE 4.-NUMBER AND PERCENT OF PERSONS WITH FEWER, SAME, OR MORE VISITS IN YEAR AFTER REFERRAL COMPARED WITH YEAR PRECEDING REFERRAL, BY TYPE OF SERVICE

6 The chi-square test was employed to test the equality of the number of patients showing

a decrease in number of visits with those showing an increase.