Older bereaved spouses' participation in self-help groups

This study reports on bereaved spouses' assessments of their participation in twenty-six self-help groups which consisted of eight weekly meetings (short-term; N = 82) or ten monthly meetings in addition to the weekly ones (long-term; N = 52). Regardless of the duration of the groups whether they were led by a professional or a widowed peer, assessments were generally positive. The most commonly reported benefit the participants hoped to gain was emotional support. No statistical differences were noted between those in the short- versus long-term condition or between those in the widow- versus professional-led groups regarding perceived helpfulness of the meetings, level of participation, or the extent to which participants' primary needs were filled. Trends, however, seemed to favor those in the long-term groups. Statistically significant correlations were observed between the number of meetings attended and a composite index of perceived value of the meetings. Major recommendations include the importance of making self-help groups that allow for greater involvement over time available for older bereaved adults, and to make use of both professionals and widowed persons as group leaders.

OMEGA, Vol. 25(1) 47-61, 1992 OLDER BEREAVED SPOUSES' PARTICIPATION IN SELF-HELP GROUPS* DALE A. LUND, Ph.D. MICHAEL S. CASERTA, Ph.D. University of Utah, Salt Lake City ABSTRACT This study reports on bereaved spouses' assessments of their participation in twenty-six self-help groups which consisted of eight weekly meetings (short­term; N = 82) or ten monthly meetings in addition to the weekly ones (long-term; N = 52). Regardless of the duration of the groups whether they were led by a professional or a widowed peer, assessments were generally positive. The most commonly reported benefit the participants hoped to gain was emotional support. No statistical differences were noted between those in the short- versus long-term condition or between those in the widow- versus professional-led groups regarding perceived helpfulness of the meetings, level of participation, or the extent to which participants' primary needs were filled. Trends, however, seemed to favor those in the long-term groups. Statis­tically significant correlations were observed between the number of meetings attended and a composite index of perceived value of the meetings. Major recommendations include the importance of making self-help groups that allow for greater involvement over time available for older bereaved adults, and to make use of both professionals and widowed persons as group leaders. This article examines findings from a study conducted to determine the effec­tiveness of self-help groups in facilitating the adjustment process of recently bereaved older spouses. The study was designed to investigate the roles of group leadership and duration by comparing self-help groups led by widows versus professionals, and short- versus long-term formats. While psychosocial 'Findings are reported from two studies funded by the National Institute on Aging (R01-AG02193 and R01-AG06244). © 1992, Baywood Publishing Co., Inc. 47 48 / LUND AND CASERTA outcomes were of major concern, the primary purpose of this article is to report on the bereaved spouses self-reported assessments of their participation in the groups, and to do so by comparing the assessments across the leadership and duration study conditions. Comparisons are made of the participants' self-reported needs, the extent to which the needs were met, the perceived helpfulness of the groups, and the level of subject participation. Another purpose is to present specific recommendations for improving the quality of future bereavement self­help groups, with particular emphasis on the leadership and duration features. Before we present the description of the study on self-help groups, it will be helpful to review selected research findings from our longitudinal study that led to the development of this second project. The first study provided a foundation for examining self-help groups in a scientifically controlled situation by documenting the need for such a study. Findings from that study have been published in several professional journals on aging, death, health, and social psychology, but par­ticularly important summaries can be found in the book Older Bereaved Spouses: Research with Practical Applications [1]. This book provides not only findings from our longitudinal study, but recent findings from several studies of spousal bereavement in the United States. LONGITUDINAL STUDY: THE FOUNDATION FOR THE SELF-HELP PROJECT Our longitudinal study included a sample of 192 bereaved spouses between the ages of fifty and ninety-three with a mean age of sixty-seven and a matched control sample of 104 nonbereaved, currently married persons who were similar to the bereaved respondents in terms of age, gender, and socio­economic status. Bereaved participants were identified through obituaries pub­lished in local newspapers, and nonbereaved persons were located through public voter registration lists. Most of the respondents were female (74%), high school graduates (78%), and Caucasian (97%). Bereaved respondents completed in-home interviews or mailed questionnaires six times over the course of two years, beginning as early as three to four weeks after the deaths of their spouses. Nonbereaved respondents completed mailed questionnaires according to the same schedule. One of the most important findings from the longitudinal study was the some­what unexpected discovery of the high degree of coping ability, resourcefulness, and resiliency among many of the bereaved spouses. We found that only about 18 percent of the bereaved spouses were having major coping difficulties two years after the death [2], and that mean scores of self-reported coping abilities were high at all six data collection periods (5.0 to 5.5 on a 1-7 rating scale) [3]. At the conclusion of the study, the respondents were asked about their willingness to participate in self-help groups if they had been available. Forty-four percent said they would have participated if they had been asked [4]. These data can beSELF-HELP GROUPS FOR OLDER BEREAVED SPOUSES / 49 interpreted to mean that there are clearly bereaved older spouses who are in need of and want to participate in self-help groups, but it is incorrect to assume that they represent the majority. Although not all bereaved spouses need self-help groups, we did find data documenting the need for self-help groups to facilitate the bereavement process for those who are most likely to have difficulties. For example, two of the best predictors of those who will cope poorly on a long-term basis are not having the opportunity to express one's thoughts and feelings with others, and not keeping busy [2]. Traditional kinds of bereavement support groups offer excellent oppor­tunities for both self-expression and keeping busy. We also found that qualitative aspects of existing social support networks can be helpful to bereaved elders, but as a single factor, networks cannot be expected to bring about positive outcomes [5]. Potential limitations in existing support networks can be partially countered by making new friends, forming relationships, and being able to confide in others while attending support group meetings. Two other findings from the longitudinal study influenced the design and sample for the self-help study that followed. First, we learned that while widows and widowers had some unique problems related to new hassles in their daily lives (e.g., widows having problems with home repairs and widowers being deficient at preparing meals), they were experiencing very similar psychosocial difficulties with similar levels of intensity [6]. Because of these commonalities, we again included men and women in the second study, and when possible, organized self-help groups that included both men and women in the same group. The respondents in our longitudinal study also suggested that it would be most helpful if the self-help groups began in the first four months of bereavement, because the early months are often the most difficult [7] and early difficulties were found to be predictive of longer-term adjustment problems [2]. The longitudinal study produced findings that were valuable to the development of the self-help intervention project. We learned about the major difficulties in the bereavement process for older spouses who were at greatest risk for poor out­comes, the limitations of social support networks that might be addressed by self-help groups, the degree of interest in support groups, and the preferred timing for the intervention. THE SELF-HELP GROUP STUDY Self-help groups have been around since the 1800's [8], and have grown in number to where there are now well over 500,000 different groups targeted for nearly every disease listed by the World Health Organization and for many other kinds of life stresses and transition periods [9]. Generally, self-help groups consist of people voluntarily meeting together for treatment or to accomplish a specific purpose [8]. Groups focus on the members' personal problems and the resolution50 / LUND AND CASERTA of mutual needs [10]. The effectiveness of self-help groups is based largely on the "helper-therapy" principle, which argues that people help themselves when they help others [11]. In the present context, it is hoped that bringing together older bereaved spouses who are experiencing similar feelings and adjustments will be beneficial for all group members, by allowing them to share successes and disap­pointments and by providing reciprocal emotional and social support. With only a few exceptions [12-15], very little rigorous scientific research has been done on the effectiveness of self-help groups. We are not aware of any controlled research study on bereavement self-help groups for older adults. Con­sidering the growing participation in self-help groups, recognizing their potential for helping bereaved persons, and being aware of the limited research on their effectiveness, we developed and completed a second study to examine self-help groups for older bereaved spouses. We organized and supervised our own self-help groups in order to maximize control over possible confounding effects, but designed the content and structure of the groups to be similar to the many bereavement support groups available in most communities. By conducting groups with an emphasis on members sharing thoughts and feelings, and providing reciprocal support rather than designing a unique intervention, we hoped to enhance the generalizability of our findings and make those findings more useful to others who offer bereavement or grief-related self-help groups. In addition to investigating the processes and impact of bereavement self-help groups, we designed the study to identify differences which might result from professional- versus widow-led and short- versus long-term groups. The profes­sional group leaders were individuals with master's level preparation and previous work experience in some type of counseling. The peer leaders were widowed females who had participated in our previous longitudinal bereavement study and had made positive adjustments to their situation. The short-term groups were conducted weekly for eight weeks, and the long-term groups met for eight weeks followed by ten additional monthly meetings. A control group that received no intervention also was utilized in the study. The focus on the type of leadership and duration of groups was selected for several reasons. First, it is important to know if bereaved elders who participate in self-help groups manage their grief in a more timely or successful manner than those who do not participate. Second, because grief is not a disease process, there may be no reason to engage profes­sionals in group work with bereaved elders if peers (other widowed persons) can do as well or better. Conversely, in some communities there may be a scarcity of widows or widowers willing to lead self-help groups and professionals may be more readily available. Third, with limited and often scarce resources, it is impor­tant to determine whether short-term interventions can be as effective as long-term conditions. Lund, Redbum, Juretich, and Caserta provide a more detailed descrip­tion of the study design, selection and training of group leaders, recruitment ofSELF-HELP GROUPS FOR OLDER BEREAVED SPOUSES / 51 subjects, and the resolution of problems encountered during the course of the study [16]. Method Potential participants were identified from obituaries in local newspapers. This method of identification was selected because other sources of mortality reporting were not readily available (only approximately 10% of deaths are not reported in the obituaries). The Utah State Department of Health, for example, reports and tabulates mortality, but their reports are often not available until three to twelve months after the death. One of the goals of this project was to contact the potential participant shortly after the death of the spouse (approximately 8 weeks). A possible sample of 1,150 bereaved spouses who met the study criteria of being recently bereaved, age fifty or over, able to speak English, and not attend­ing another support group was identified. Only 30 percent (N = 339) agreed to participate and completed the T1 (baseline) questionnaires (241 were assigned to be in one of the self-help groups, and 98 were assigned to be in the control condition and not attend a self-help group). As expected, the accep­tance rate for the controls (38.0%) was higher than for those assigned to the self-help conditions (27.0%). The most common reasons for refusal to partici­pate were that the subjects were not interested (60.2%), too ill (17.3%), too busy (11.8%), or too upset (3.1%). Other reasons for non-participation were unclear. Because of the delay between the completion of the T1 data and the beginning of some of the self-help groups, sixty-six bereaved spouses decided to discontinue their participation, and therefore did not attend support group meetings. Of the ninety-eight controls, thirty-six were men (36.7%) and sixty-two were women (63.3%). Their ages ranged from fifty to eighty-eight years (M = 67.6 years, SD = 8.7). Similary, of the 175 who attended at least one self-help meeting, there were forty-two men (24.0%) and 133 women (76.0%) with an age range from fifty to eighty-nine years (M = 67.3 years, SD = 8.2). A telephone survey of 125 randomly selected persons who refused to participate conducted one year later yielded no significant demographic differences between participants and non-participants. Those who had refused, however, reported slightly lower stress levels and fewer problems with loneliness, but also lower perceived coping abilities and lower self-reported health ratings. Twenty-seven groups were formed; fourteen were professionally led and thirteen were widow led. One of the professional groups met only once. Subsequently, the members decided not to attend additional meetings because of inconvenience and lack of interest. Eighteen of the groups had male members, and only six groups had just one male member. The average size of the groups was approximately six members. Several leaders, both peer and professional, led more than one group, but no one led more than three groups.52 / LUND AND CASERTA Procedure Data were collected from personally delivered questionnaires (Tl) or mailed questionnaires (T2 through T4) at four points in time, following the same schedule for all participants regardless of their group assignment. The first questionnaire was delivered at approximately two months after the death of the spouse, and mailed back prior to attendance at support group meetings. The personal visit at Tl was made to further explain the study and encourage continued partici­pation. The T2 measures were obtained immediately after the completion of the eight weekly self-help meetings (roughly 4-8 months bereaved). The T3 questionnaires followed the completion of the long-term groups (roughly 14-17 months bereaved), and T4 was obtained at two years following the death. The control group also was assessed at the same four time periods, but received no intervention. The questionnaires included items regarding demographics, socioeconomic status, competencies, self-esteem, and social support factors. In addition, the principal outcome measures included instruments to assess grief resolution, depression, life satisfaction, and perceived coping, stress, and health. These out­come measures are the focus of another article which is currently being updated and revised and is not part of this report [17]. The measures used in the present analyses deal with the participants' reported needs, their level of participation in the groups, and the perceived helpfulness of the groups. The bereaved spouses were asked to select, from a list of five choices, "the main need that you hoped would be filled by participating in these groups." The answer choices included "receive emotional support," "obtain information about doing certain tasks," "meet new people," "desire to help others," and an option to specify something else not on the list. The next question asked the participants to indicate on a scale from 1 (not at all) to 5 (very well) "how well their need was filled by par­ticipating in the group sessions". On a similar 1 (no at all helpful) to 5 (extremely helpful) rating scale, they were also asked, "How helpful were these meetings for you?" and "How would you characterize your level of participation in the group meetings?" Reclassifying Participants into Short-and Long-Term Conditions , / Although respondents were assigned to a specific treatment condition, some did not attend enough meetings to be considered as valid participants in that condition. We determined that the minimum requirement for inclusion in the short-term condition was attendance at least one-half (four) of the weekly meetings. Likewise, the criterion for being a valid long-term participant was to have attended at least one-half (five) of the monthly meetings, in addition to atSELF-HELP GROUPS FOR OLDER BEREAVED SPOUSES / 53 least four of the weekly meetings. It was therefore possible, for purposes of categorical-level statistical analyses, for a respondent to cross conditions by not meeting the established criteria. A respondent in either of the treatment groups could have crossed into the control condition by not attending any meetings but still completing questionnaires throughout the study. Similarly, a respondent who was assigned to a long-term condition may have attended enough weekly meet­ings to fulfill the short-term criteria but failed to attend any monthly meetings. Therefore, the total sample (N = 134) for the categorical analyses consisted of eighty-two respondents who met the minimum criteria for the short-term condi­tion, and fifty-two in the long-term condition. Those who met the criteria for the short-term condition attended a mean of 6.5 (SD = 1.3) meetings, compared with a mean of 15.2 (SD = 2.0) total meetings for those in the long-term condition. No statistical differences in attendance were detected between the widow- versus professional-led groups. Fifty-six percent of those in the short-term condition attended at least seven of the eight meetings, with 27 percent having perfect attendance records. Fifty-four percent of those in the long-term condition attended at least sixteen of the eighteen meetings, with 6 percent having maximum atten­dance records. Because self-selection was a concern, statistical tests were performed on both demographic variables and baseline outcomes comparing those who remained "true" to their assigned condition with those who were treated as crossing over to other conditions. No statistical differences were found between assigned controls and those in the treatment conditions who attended no meetings. Also, no statis­tical differences were found between those who met the long-term criteria and those who did not in that condition, or between those who attended at least four weekly meetings and those who attended none within the short-term condition. We recognize that self-selection into the study design and the crossing over into another condition are limitations of the study, but we believe that the advantages of these procedures outweigh the liabilities. There were also forty-one respondents who can best be classified as partial participants. These were bereaved spouses who attended at least one meeting but did not fulfill the minimum criteria for any treatment condition. The average number of meetings attended by these forty-one partial participants was four (SD = 3.5). Two-thirds of them attended three or fewer meetings. Data from these partial participants were used only in the correlational analyses dealing with the number of meetings attended. Seventy-six percent of the 134 participants (82 short-term and 52 long-term) described above were women, with over 90 percent having graduated high school and 69 percent having an annual household income of less than $20,000. The average age of the sample was 67.6 years (SD = 8.1) and the participants had been married an average of 40.2 years (SD = 13.7) at the time of their spouses' deaths. Sixty-four (47.8%) were in a group led by a professional, and seventy (52.2%)54 / LUND AND CASERTA were in a group led by a widowed peer. No statistically significant differences were observed on these demographic characteristics with respect to the type of group (duration and leader) in which subjects participated. RESULTS As we indicated earlier, we have neither focused attention in this article on the affective and behavioral outcomes of the group meetings, nor made comparisons with those in the control group condition. Our preliminary analyses on these outcomes have revealed little or no differences between the bereaved persons in the various study conditions [17], but more extensive analyses are required to confirm, clarify, and explain the lack of significance. An equally important part of this study, however, was to examine the bereaved persons' reasons for participat­ing, the extent to which their needs were met, and their perceptions of their participation in and the helpfulness of the group meetings. Analyses of these data will help to develop more satisfying and effective bereavement self-help groups, particularly for older adults. As expected, no differences were found between the bereaved spouses in any of the self-help conditions (long- or short-term; widow- or professionally led) regarding the primary need that they hoped would be filled by participating in the groups. Their perceived needs were expected to be independent of their assigned study condition, and the chi-square analyses provided confirmation. Table 1, however, reveals that a clear majority of those in our study wanted foremost to receive emotional support. Nearly 61 percent of the bereaved spouses chose this as their main need. About 17 percent indicated that their primary reason for participating was a desire to help others, followed by 6 percent who wanted to obtain information about completing tasks. With respect to the bereaved persons' perceptions of how well their main need was filled by their participation, there were no significant differences attributed to the type of self-help group condition. The same questions were asked of the participants at the end of the eight weekly sessions (T2), at the conclusion of the ten monthly sessions (T3), and two years after the death event (T4) to determine if their perceptions of their involvement changed over time. The mean scores on a scale from 1-5 are shown in Table 2. Although no statistical differences were found, it should be noted that the mean scores appear to be moderately high (ranging from 3.57 to 4.26) and slightly more favorable for those in the long-term condition (regardless of whether it was a widow- or professional-led group). Caution must be exercised not to attribute statistical significance, but all of the mean scores of need fulfillment were higher for the long-term group condition than they were for the short-term condition. Also, the mean scores remained quite stable over time, which suggests that the perceived level of need fulfillment remained strong even with the passage of time.SELF-HELP GROUPS FOR OLDER BEREAVED SPOUSES / 55 Table 1. Major Perceived Need According to Self-Help Group Format3 Major Perceived Need (at Time 2) Long-Term Short Term Widow Led (N = 30) Professional Led (A/ = 21) Widow Led (N = 35) Professional Led (N = 39) Emotional support 53.3% 66.7% 62.9% 61.5% (16) (14) (22) (24) Desire to help others 13.3% 23.8% 14.3% 17.9% (4) (5) (5) (7) Meet new people 16.7% 9.5% 5.7% 17.9% (5) (2) (2) (7) Obtain information 6.7% 0.0% 11.4% 2.6% (2) (0) (4) (1) Other need 10.0% 0.0% 5.7% 0.0% (3) (0) (2) (0) "Respondents asked to select only one need rather than checking all of them. Chi-square = 13.48 (12 df)\ N.S. Again, regarding the participants' perceptions of the helpfulness of the support group meetings, we found no statistically significant differences among the types of self-help group conditions. The mean scores presented in Table 3, however, reveal a slight advantage to those in the long-term support groups, with all of their mean scores (T2, T3, and T4) higher than those of the short-term groups. These data parallel those found for need fulfillment in terms of the generally positive assessment of the groups (mean scores ranging from 3.68 to 4.24) and stability of the perceptions over time. Because the results were strikingly similar regarding the needs being filled by group participation and the perceived helpfulness of the groups, and also because there appeared to be a slight advantage for those in the long-term condition, we computed a summed score of the two items and conducted a t-test comparing the mean scores for those in the short- versus long-term groups. The possible scores on the combined index which we refer to as the participants' perceived value of the group meetings could range from 2 to 10. The Cronbach alpha coefficient for the index was .80. The results of the t-test showed a borderline significant dif­ference (p < .06), with those in the long-term condition having more favorable ratings at both T2 (A/ = 8.2; SD = 1.7) and T4 (M = 8.1; SD = 1.8) than those in the56 / LUND AND CASERTA Table 2. Extent to Which Major Perceived Need was Filled by Self-Help Group Format How Well Was Major Need Filled by Participating8 Long-Term Short Term Widow Led Professional Led Widow Led Professional Led Tinr^ M 3.97 4.24 3.69 3.79 (end of 8 weeks) (SD) (1.00) (0.94) (1.02) (1.04) N 30 21 35 38 Time 3C M 4.14 4.26 _ _ (end of long-term) (SD) (0.88) (1.05) (not asked) (not asked) N 29 19 Time 4d M 4.00 4.11 3.57 3.60 (2 yrs. post-death (SD) (0.83) (0.94) (0.92) (1.24) event) N 27 19 28 35 a1 = Not at all; 5 = very well. bF (3,120) = 1.48, N.S. CF (3,146) =0.20, N.S. dF (3,105) = 1.82, N.S. short-term groups (T2:A/ = 7.6, SD = 1.8; T4:A/ = 7.4, SD = 1.9). The t-value at T2 was -1.86, and at T4 it was -1.88. In order to test for the strength of the relationships between the number of support group meetings attended and the composite index of perceived value of the groups, we computed Pearson correlations at both T2 and T4. These analyses have the advantage of testing for a continuous linear relationship between the variables, rather than limiting the tests at the categorical level of short- versus long-term groups and ignoring the actual number of meetings attended. For these correlations, we included the forty-one partial partic ipants who attended at least one meeting but did not meet the 50 percent attendance record required to be included in the short- versus long-terms comparisons. At both T2 and T4, the number of meetings attended was moderately associated with the perceived value of the groups. Those who attended more meetings reported the meetings as being more valuable (r = .30, p < .0001 and r = .29, p < .001, respectively). We computed the same correlations while excluding the partial participants, and the strength of association dropped slightly to r = .21 (p < .02) and r = .22(p< .02) at T2 and T4 but remained statistically significant. Although these correlations can be interpreted to mean that bereaved spouses who perceived the meetings as being less helpful and not meeting their major need decided to attend fewer meetings, it is equally plausible that the value of the self-help groups cannot be fully realizedSELF-HELP GROUPS FOR OLDER BEREAVED SPOUSES / 57 Table 3. Perceived Helpfulness of Meetings According to Self-Help Group Format Perceived Helpfulness of Meetings'* Long-Term Short Term Widow Led Professional Led Widow Led Professional Led Time & M 4.20 4.00 3.74 3.92 (end of 8 weeks) (SD) (0.92) (0.97) (0.92) (0.90) N 30 20 35 39 Time 3C M 4.24 4.21 3.70 3.77 (end of long-term) (SD) (0.87) (1.08) (1.10) (1.03) N 29 19 27 35 Time 4d M 4.11 3.95 3.68 3.83 (2 yrs. post-death (SD) (0.89) (1.13) (1.12) (0.89) event) N 27 19 28 35 *1 = Not at all helpful; 5 = extremely helpful. bF (3,120) = 1.35, N.S. CF (3,106) =2.08, N.S. dF (3,105) = 0.92, N.S. unless the participants give them the opportunity to be effective by attending most of the meetings. We have evidence which shows that the partial participants also had relatively positive feelings about the support groups. For example, their mean scores on the 1-5 rating scales for need fulfillment and perceived helpfulness were both 3.2, and on a similar single-item measure of their reported level of participation during the group meetings, their mean score was 3.8. Also, based on the exit surveys com­pleted by the partial participants who notified us about discontinuing their atten­dance at the meetings, we learned that some had moved away from the local area and others had changes in their employment or personal lives which made their attendance either inconvenient or impossible. The final item concerning the bereaved adults' assessment of their group experiences examined their reported levels of participation during the group meeiings. In order to maximize the effectiveness of support groups, it has been generally assumed that active participation is required of the members. This single-item measure, although limited in scope, provides some basis to test the assumption about the importance of active participation and whether it varied across the types of self-help groups. At T2, which corresponds with the comple­tion of the eight weekly sessions, there were no significant differences among the leadership and duration conditions. The mean scores were moderately high and58 / LUND AND CASERTA ranged from 3.69 to 4.05. At T3, however, there was a borderline effect (p = .09) within the long-term group participants, which revealed slightly higher levels of participation by those attending groups led by professionals (M = 4.32, SD = .67) rather than those led by the widowed peers (M = 3.90, SD = .90). It could be that the professional leaders were more skilled at facilitating this aspect of group process over a longer period of time. It would be inappropriate to assign more importance to this issue, given the limited information and borderline nature of the effect. However, for those who might expect professional leaders to do too much talking and thereby limit the active participation of the bereaved group members, our data would refute the assumption. It is also worthwhile to note that after the eight weekly meetings we found no significant relationship (r - .12, N.S.) between the number of meetings attended and the participants' reported level of involvement at the meetings. However, after the completion of the ten additional monthly meetings, the Pearson correlation between these same two items was .53 (p < .0001). These correlations can be interpreted to mean that active participation of the members at support group meetings may not be necessary for their attendance if they are expected to attend only eight meetings, but is quite important to ensure the members' continued attendance over a much longer period of time. Another possibility is that as members attended more meetings, they perceived themselves investing a greater amount of time and energy into the meetings, and therefore participated at a higher level. CONCLUSIONS AND DISCUSSION Although the findings presented in this article are limited to specific aspects of bereavement self-help groups for older adults, we have learned several important and useful pieces of information. In summary, we would like to highlight five specific findings. First, the older bereaved spouses in our study reported relatively positive ratings about their participation in the self-help groups regird less of the type of group leadership and duration. As a full sample, they gave positive assessments about their primary needs being met, their perceived helpfulness of the groups, and their levels of attendance and participation at the meetings. Second, it was very clear that receiving emotional support was the most common need identified by the bereaved spouses. A smaller proportion of them mentioned either a desire to help others or to meet new people as the need they wanted filled by their participation in the self-help groups. Third, the type of group leader did not significantly impact either the bereaved group members' abilities to fill their primary needs or their assessments of the helpfulness of the groups. One slight advantage was found for the long-term professional-led groups, where the bereaved participants appeared to reportSELF-HELP GROUPS FOR OLDER BEREAVED SPOUSES / 59 slightly higher levels of participation during meetings than did their counterparts in the widow-led groups. Fourth, a pattern emerged in the data that might suggest a slight advantage to those who participated in the long-term groups. Although the tests for statistical significance were inconclusive, there was evidence suggesting that those who attended more meetings over a longer period of time reported that the groups were more helpful, and that their primary needs were met more fully. We suggest that the potential value of self-help groups requires attendance at more than just a few meetings. Fifth, if support groups are expected to continue beyond eight consecutive weekly meetings, the group members' subsequent attendance will depend upon their perceived active involvement at the meetings. It is quite likely that those who do not actively participate will attend fewer meetings over the course of several months. Based upon findings from our self-help group study and, to some extent, the results of our previous longitudinal study, we believe that several specific recom­mendations should be made regarding future bereavement self-help groups. First, it should not be necessary to state, but we want to underscore the importance of recognizing that older adults have needs similar to those of younger people. It is equally important to older adults to have self-help groups made available to them, whether it be for purposes of bereavement or the many other situations where reciprocal sharing and support might be useful. We cannot neglect the needs of older adults based on past misconceptions and stereotypes. Second, we believe that men and women have many similar experiences and feelings during bereavement, and although it might be more difficult to get men to participate in bereavement groups, we need to continue our efforts to solicit their participation by offering groups which might include men and women together in the same groups. Third, we highly recommend using both professional and widowed persons to lead bereavement support groups. When organizing these groups, one must assess the human resources available in the local communities and make use of the most convenient and cost-effective strategies. Professionals and older widowed persons are capable of yielding similar effective results. Fourth, we recommend that bereaved older adults be given the opportunity to participate in self-help groups which last longer than eight weekly meetings. The perceived value to the participants is likely to increase when meetings can con­tinue for up to one year. Obviously, not all older bereaved adults will need or want to participate in self-help groups, or to do so for as long as twelve months, but we should strive to offer the opportunities with the flexibility necessary to maximize the potential value of the groups. Finally, because of the many unexpected findings that we have encountered from our research on bereavement over the past eleven years, we strongly encourage other controlled studies of bereavement, and particularly studies of60 / LUND AND CASERTA self-help groups. We are only beginning to scratch the surface of what we need to know to have self-help groups realize their potential. r ACKNOWLEDGMENT The authors wish to acknowledge Dr. Margaret Dimond who served as a co-investigator of the study and thank the anonymous reviewers of the article for their thoughtful suggestions. REFERENCES 1. D. A. Lund, Older Bereaved Spouses: Research with Practical Applications, Taylor & Francis/Hemisphere, New York, 1989. 2. D. A. Lund, M. F. Dimond, M. S. Caserta, R. J. Johnson, J. L. Poulton, and J. R. Connelly, Identifying Elderly with Coping Difficulties after Two Years of Bereave­ment, Omega, Journal of Death and Dying, 16, pp. 213-224,1985. 3. D. A. Lund, Conclusions about Bereavement in Later Life and Implications for Interventions and Future Research, in Older Bereaved Spouses: Research with Prac­tical Applications, D. A. Lund (ed.), Taylor & Francis/Hemisphere, New York, pp. 215-231,1989. 4. D. A. Lund, M. F. Dimond, and M. S. Juretich, Bereavement Support Groups for the Elderly: Characteristics of Potential Participants, Death Studies, 9, pp. 309-321, 1985. 5. M. F. Dimond, D. A. Lund, and M. S. Caserta, The Role of Social Support in the First Two Years of Bereavement in an Elderly Sample, The Gerontologist, 27, pp. 599-604, 1987. 6. D. A. Lund, M. S. Caserta, and M. F. Dimond, Gender Differences through Two Years of Bereavement Among the Elderly, The Gerontologist, 26, pp. 314-320,1986. 7. D. A. Lund, M. S. Caserta, and M. F. Dimond, Impact of Spousal Bereavement on the Subjective Well-being of Older Adults, in Older Bereaved Spouses: Research with Practical Applications, D. A. Lund (ed.), Taylor & Francis/Hemisphere, New York, pp. 3-15,1989. 8. A. H. Katz and E. I. Bender, The Strength In Us: Self-help Groups in the Modern World, New Viewpoints, New York, 1976. ' 9. A. J. Gartner and F. Riessman, Self-help and Mental Health, Hospital and Community Psychiatry, 33, pp. 631-635,1982. 10. A. Withom, Helping Ourselves: The Limits and Potential of Self-help, Social Policy, 11, pp. 20-27,1980. 11. F. Riessman, The "Helper-Therapy" Principle, Social Work, 10, pp. 27-32,1965. 12. M. L. S. Vachon, W. A. L. Lyall, J. Rogers, K. Freedman-Letofsky, and S. J. J. Freeman, A Controlled Study of Self-help Intervention for Widows, American Journal of Psychiatry, 137, pp. 1380-1384,1980. 13. R. Constantino, Bereavement Crisis Intervention for Widows, Nursing Research, 30, pp. 351-535,1981.SELF-HELP GROUPS FOR OLDER BEREAVED SPOUSES / 61 14. M. A. Lieberman and L. D. Borman, The Impact of Self-help Groups on Widows' Mental Health, National Reporter, 4, pp. 2-6,1981. 15. M. A. Lieberman and L. Videka-Sherman, The Impact of Self-help Groups on the Mental Health of Widows and Widowers, American Journal of Orthopsychiatry, 56, pp. 435-449,1986. 16. D. A. Lund, D. E. Redbum, M. S. Juretich, and M. S. Caserta, Resolving Problems Implementing Bereavement Self-help Groups, in Older Bereaved Spouses: Research with Practical Applications, D. A. Lund (ed.), Taylor & Francis/Hemisphere, New York, pp. 203-216,1989. 17. D. A. Lund, M. S. Caserta, and M. F. Dimond, Effectiveness of Self-help Groups for Older Bereaved Spouses, paper presented at the 42nd Annual Meeting of the Geron­tological Society of America, Minneapolis, Minnesota, November 1989. Direct reprint requests to: Dale A. Lund, Ph.D. University of Utah Gerontology Center 25 South Medical Drive Salt Lake City, UT 84112