role of support networks in maintenance of improved cardiovascular health status

Sot. Sci. Med. Vol. 28. No. 3, Pp. 249-260, 1989 Printed in Great Britain. All rights reserved

0277~9536189 S3.00 + 0.00 Copyright ,c 1989 Pergamon Press plc

ROLE OF SUPPORT NETWORKS IN MAINTENANCE OF IMPROVED CARDIOVASCULAR

HEALTH STATUS

PATRICK O’REILLY’ and H. EMERSON THOMAS*

‘Cardiovascular Institute, Boston University School of Medicine and lSection of Preventive Medicine and Epidemiology, Department of Medicine, Boston University School of Medicine, 80 East Concord Street,

Boston, MA 02118, U.S.A

Abstract-This study was undertaken to clarify the relationship between maintaining an improved cardiovascular health status and social support networks. Two hundred and ninety participants from a national trial that was carried out to specify the impact of reducing risk for cardiovascular disease were rescreened 3 years after completion of the trial. Original risk status of the participants was compared to risk status at the end to identify who had an improved risk status (n = 204). Risk status of 204 improvers was calculated from the rescreening data to identify maintainers (n = 63) and nonmaintainers (n = 143).

No significant differences were found betwen maintainers and nonmaintainers in sociodemographic status, or in level of general support. Highly significant differences were found for four types of support provided specifically for risk reduction: information/advice (P = 0.002), appraisal (P = 0.004), emotional support (P = 0.01) and availability (P = 0.019). Most of these differences in support were accounted for by the larger support network identified by the maintainers. In addition, compared to nonmaintainers, maintainers’ networks were more family-centered (P = 0.012). and were correspondingly more dense (P = 0.021). A discriminant function analysis using the significant variables was able to predict maintenance in 72% of the cases. The significant variables also accounted for 10% of the variance between maintainers and nonmaintainers.

The implications of this study are: (1) assessment of support should be specific to the health outcome being studied; (2) research on the impact of social environment on health status will benefit from clearly specifying individual components of social support and social networks; (3) amount of social support provided is related to size of the network and relationship of supporter to the at-risk individual; (4) because social support and social networks account for a small but significant amount of the variance between maintainers and nonmaintainers, these concepts should be included in patient assessments prior to developing a treatment or maintenance plan.

INTRODUCTION

In the past two decades, the United States has witnessed a dramatic decline in the mortality rate for coronary heart disease [l]. Part of the cause of this decline can be attributed to the development of more precise surgical and diagnostic techniques and to newer and better pharmacological treatment. How- ever, this decline also is in no small measure due to the identification of cardiovascular risk factors, prin- cipally hypertension, cigarette smoking, and high cholesterol levels, and to the professional and public activities directed at reducing the incidence and prevalence of these risk factors.

However, while numerous individual-, group-, and community-based studies have been found to be successful at lowering risk for heart disease [2-71, these programs have also experienced exceptionally high rates of recidivism. Early gains have not ensured long-term success, and heart disease and other modifiable health problems continue to place an unwarranted and unnecessary burden on the health- care delivery system.

Investigators looking into ways to attain and maintain changes in personal health behaviors have begun to provide evidence of the role that social support and/or social network members have in changing such personal health behaviors as smoking [8,9] and weight control [ 10, 11). Other studies have focused on

the role of support or supporters in complying with medical regimens for controlling such problems as high blood pressure [12, 131, diabetes [14], and asthma [15, 161. However, for most of these studies support was not measured. Rather, it was assumed to be provided by the inclusion of home visits, significant other training, structural reinforcement and/or support groups in the intervention design. In a review of a number of these studies, Levy [17] was led to conclude that few of them presented a clear view of what support was provided, thereby making it difficult to interpret the effect of support on compliance. Most of these studies also did not collect data on the relationship of support to long-term maintenance of the changes in the study population.

A notable exception is the work of Mermelstein el al. [18] on smoking cessation. The authors measured both the perceived availability of general support and support directly related to smoking, as well as the relationship of these two types of support to maintenance 3 and 12 months after quitting. Neither type of support was found to differentiate abstainers from recidivists.

Extending this approach, the present study focuses on the issue of the relationship of social support and social networks to maintenance of improved cardiovascular health status. Measures of general social support and support specific to reducing risk for heart disease were used, along with a measure of the

249

250 PATRICK O’REILLY and H. EMERYDN THOMAS

close social network. The use of these different Using this procedure, the risk scores at Yr, were measures enabled us to note separately what effect compared to the risk scores at Yr, for all 290 different types of support and different components rescreening participants. Two hundred and four par- of the close network had on long-term maintenance. ticipants (70%) had risk scores lower at Yr, than at

Yr,. These individuals were classified as ‘improvers’

METHODOLOGY and comprised the population for this study.

This study was carried out as part of the rescreening of a population of middle-aged men who partici-

Physiological &rc

pated in a national study called the Multiple Risk Upon arrival at the rescreening site, measurements

Factor Intervention Trial (MRFIT) [19]. of height, weight and blood pressure were taken

The aim of this trial was to clarify what effect along with the drawing of a sample of blood. The

reducing the major risks of cardiovascular disease procedure for the collection of these data was consis-

(i.e. hypertension, high cholesterol, and cigarette tent with that employed during the MRFIT trial,

smoking) had on mortality rates. Boston University including the appropriate preparation and storage of

School of Medicine was one of 22 sites that carried blood specimens before shipment to the MRFIT

out this project between 1976 and 1982. The rescreen- Coordinating Center laboratory for chemical deter-

ing was designed to assess the cardiovascular health mination. Self reports on smoking and the use of

status of the MRFIT population 3 years after the risk-factor-related health and medical services were

original study was completed. Contact had not been also collected.

mamtained with the study population in the interven- ine vears. Social support and social network measures

‘ibe rescreening population was a selective sample drawn by the MRFIT Coordinating Center from the original 630 participants in the Boston University MRFIT program. Funding and time permitted the selection of 465 of these individuals for inclusion in the rescreening. Upon attempting to contact these individuals, it was found that 61 had either dropped out before the end of the study, had died since the end of the study, or were too seriously ill to participate in the rescreening. This process left 404 participants eligible and available for inclusion in the MRFIT rescreening. Of this number, 290, or 72% agreed to participate.

Because the primary aim of this study was to clarify the role of social support and social networks in the maintenance of an improved health status, the study population included only those rescreening participants who had completed the program and who, on completion, had changed any of their risk factors such that their cardiovascular risk status had improved. To determine how many participants had improved and how many had maintained the im- provement, it was necessary to calculate the participants’ risk status at three points of time: (1) at entry into the program (Yr,); (2) at the end of the program (Yr6); and (3) at the time of the rescreening (Yr,). Risk status was calculated using a logistic equation based on data from the Framingham Heart Study, which uses data on factors known to be related to the risk of developing coronary heart disease [20]. These factors are: sex, age, blood pressure, serum cholesterol, cigarette smoking, blood sugar. and ECG abnormalities. Lacking information on any one of these factors, that factor is assumed to be normal. For this study the five risk factors used to establish the risk index were: age, sex, blood pressure, serum cholesterol, and cigarette smoking. Data were not available on blood sugar levels and ECG abnormalities. Because age has such a powerful influence on establishing risk and is not a factor which can be modified, it was decided to control for age by using the participants’ age at Yr, as a constant when the two additional risk indexes were calculated.

Methodological reviews of instruments designed to measure social support and social networks [2O-251 have found conceptual and operational ambiguity and a lack of distinction between components of social support and social networks. In addition to definitional issues, some instruments, while found to be conceptually valid and operationally reliable, were overly lengthy in their administration and/or were not specific in the support or network variables utilized. Finally, none of the instruments reviewed clearly provided an option not to seek or accept support (e.g. a network orientation). Therefore, it was decided to utilize the best features of some of the existing instruments and to add sections to complete the information necessary to obtain data on both general support, and support for risk-reduction efforts.

For this study, social support was viewed as an interactional process in which particular actions or behaviors directed at an individual have a positive effect on that individual’s social, psychological or physicial well being. Included in the operationalization of this definition were behaviors related to instrumental, emotional, appraisal, information/ advice and confidant support. The questions were selected to reflect the diverse functions of social support, and what the literature suggests as the principal categories in which these functions can be placed [23,25]. Also, the five questions adapted from other sources follow the categorization of the original sources.

The format for the first questionnaire, which was mailed to the participants prior to their MRFIT rescreening, was adapted from the format used by Henderson et al. (261, Schaefer et al. [27], Barrera [28], and Fischer [29]. These investigators asked respondents from whom had/would they receive help in certain common situations (i.e. looking after a house, needing advice, having a family problem). This format is effective in identifying the presence of support and who the providers are. However, it does not directly ask individuals if in these situations they would choose to handle the situation themselves; that is, not ask or look for help or support. The latter

attribute can be considered an indicator of an individual’s network orientation, and it was decided to include it as one of the principal response categories for the situations described in the questionnaire.

Based on the work of the authors cited above, as well as the theoretical and empirical issues in the literature, it was decided to elicit responses about seven situations. Each of the situations would be an indicator of a specific type of social support that could be offered in that situation and who, if anyone, provided that support. The seven situations, the source of the question, and the type of support to which each is related are listed in Table 1, below.

While specifying a particular type of support as being related to each of the seven situations used in the questionnaire, we appreciate that categories of support cannot be so easily separated; one type of support (e.g. instrumental) could be interpreted by the recipient of that behavior as an indicator of another type of support (e.g. encouragement). Since the question of how the recipient interpreted any of the supportive situations was not asked, it will be necessary to interpret their responses at face value.

Two additional questions were included. First, on a scale of l-5, with responses ranging from very dissatisfied to very satisfied, respondents were asked to state the degree to which they were satisfied with each of the situations; that is, with or without obtaining help from others. A second question asked respondents, irrespective of the degree of satisfaction, if they would like to have more help/support, less help/support, or about the same amount. The format for the latter question allows information to be obtained about whether or not the provision of too little or too much help/support is of concern to the respondents.

Finally, because participation in social activities often is interpreted as a form of support or of network orientation [30], an eighth question, using the format described above, asked the respondents how they usually carried out social activities, how satisfied they were with this situation, and whether or not they would like to see any changes.

The final format, then, for the general support questionnaire consisted of a series of questions, each

Role of support networks 251

with two principal parts. The first part consisted of seven briefly-described common situations in which respondents were asked if they would usually: (1) ask for or obtain help/support from others in the situations described, and if so, from whom; (2) handle the situations without asking or expecting help/support; or (3) choose some other way to handle the situations. Using the same format an eighth question asked about socializing with others. (General support questions are listed in Appendix.) The second part asked respondents the degree of satisfaction experienced in the eight situations, and whether they would like any to change (i.e. get more or less help/support) or remain as is. The second questionnaire, completed as part of the personal interviews at the time of the MRFIT rescreening, also had two parts. The first part was adapted from Schaefer et al. [27] and was designed to measure support specific to risk- reduction efforts. Respondents were first asked to identify, from the list of individuals named in response to the general support questionnaire, up to five individuals from whom they had received the most help, assistance, or reassurance. Provision was made to allow participants to respond with ‘no one’ or to add the names of up to three others not previously identified but to whom they felt particularly close or who had been particularly helpful. Names elicited in this manner were used only if five individuals were not identified in the first questionnaire.

For each of the individuals named, four questions related to help or concern with risk-reduction efforts and health were asked. For each question, the participants were asked to rate, on a scale of one to five, the help or support that each individual provided, from little or no help to extremely helpful. The four areas of support were: (1) information and advice; (2) appraisal; (3) emotional support; and (4) availability. Two additional questions using the same format were also included. One asked the degree of satisfaction with the support given by each individual and the second, the degree of conflict respondents had with each of these individuals. Interviewers informed participants that conflict did not refer to minor day-to- day disagreements, but to difficulties that had

Table I. General support questions: source and type of support

Sttuation SOtIKe Type of support

I.

2.

3.

4.

5.

6.

7.

Ask someone if you could borrow a

tool or some equipment to finish a

job around the house

Ask someone to take care of house,

pets

Seek advice when making important

decisions

Confide in others when worried or

concerned about a personal matter

Be shown cmcern when gomg through

a particularly difficult ttme

Be given encouragement in carrying

out different activities

Be given help/encouragement in

Henderson (261

Schaefer [27]

Instrumental

New questmn Instrumental

Barrera [28]

Fischer [29]

Henderson [26]

Schaefer [27]

Advtceiopinion

Barrera [ZS] Confidant

Fischer [29]

Henderson 1261

Schaefer ]2?] 1 Henderson [26] Emotional

Barrera [28] Encouragement;

Henderson [26] appraisal

New questmn Emotional/

managing personal health problems encouragement

252 PATRICK O'REILLY and H. EMERSON THOMAS

changed or affected their relationship with the individual. (Support questions specific to risk-reduction are listed in Appendix.)

The second section of the second questionnaire focused on specific dimensions of the social network. A personal social network can be defined broadly as those individuals to whom one has direct links. or it can be restricted to the close support network (i.e. significant or important ties that provide support to an individual). For two reasons, this study has focused on the latter-the more narrow---close support network: (1) there is some justification for stating that close family and friends are likely to have the most positive or negative effect upon a person’s personal health behaviors [31, 321; (2) a number of studies that have looked at the relationship between health and social networks have focused on the close personal network of the focal individuals [33-361. By limiting this study to a similar network, it should be possible to make some broad comparisons between previous studies and this one. As for the specific dimensions of the network, it was decided to include those dimensions that either the literature indicated could have an effect on health or that seemed to be logically con- nected to personal health habits. Those social network dimensions with their operational definitions follow:

I. 2.

3. 4.

5. 6.

7.

Size: number of persons in close network. Content: domain of relationship to focal individual (i.e. spouse, friend). Gender: male or female. Distance: geographical distance in terms of time it takes for the focal individual to reach a network member (i.e. less than 5 min, 5-60 min. more than 60 min). Duration: length of time of the relationship. Frequency of contact: amount of contacts between focal individual and network members (i.e. daily, weekly, monthly). Density: proportion of individuals in the close network who know one another.

In summary, the format for the second questionnaire had two parts, with both parts obtaining information about previously-identified. supportive individuals. The first series of questions concerned the type and degree of help/support these individuals had provided to participants in their risk-reduction efforts. how satisfied they were with the support; help provided, and ended with a final question on the degree of conflict between participants and individuals named in the network. The second series of questions obtained information on the structural and interactive dimensions of the close support network.

*From data collected at the end of the original MRFIT program. nonparticipants were found to have the same high marriage rate (97%) and were slightly younger (mean age = 54) than participants. These data also showed no differences between participants and nonparticipants in diastolic blood pressure (82.1 vs 83.2), cholesterol level (235.0 vs 235. I) and percentage of those who quit smoking (37.0 vs 27.4; P = 0.17).

RESULTS

An analysis of MRFIT rescreening data revealed that 63 of 204 improvers, who comprised the study population, had maintained the improved risk status achieved by the end of the original MRFIT study. That is, after controlling for age, the risk status of 63 participants 3 years after the end of MRFIT was equal to or lower than their risk status at the end of MRFIT. This group will be referred to as the maintainers. The I41 improvers who did not maintain the improved risk status achieved at the end of MRFIT will be referred to as the nonmaintainers.

Sociodemographic status

No significant sociodemographic differences ex- isted between the two groups. The population was middle-aged and older (average age 55, range 4567). Over 90% were married. They were predominantly Catholic (67%) and from a large variety of ethnic backgrounds, particularly Irish (36%), English (17%). and Italian (10%). A total of 28 different ethnic groups were identified by the participants. The population contained an occupational mix with more participants identified as white collar (53%) than blue collar (39%) or retirees (18%).*

General social support

As seen on Table 2, no significant differences were found between the percentage of maintainers and nonmaintainers who usually asked for or expected to receive help/support in the seven situations provided. Nor was there any difference between the two groups in the percentage of participants who socialized with others. The data also show that for both groups, participants were most likely to socialize with others and least likely to borrow something. In addition, approximately one-third in both groups usually man- aged common health problems by themselves.

No differences were found between maintainers and nonmaintainers in the choices of individuals from whom they usually asked for or expected to get help/support in the seven supportive situations and in the socializing with others question (data not shown). Participants in both groups selected wives most frequently for all the support variables, except for borrowing and taking care of house. Other relatives and neighbors were most often selected for these two types of instrumental support. Wives were also the participants’ first choice for the person with whom they would like to socialize. In looking at all the choices for supporters, maintainers

Table 2. DilTerences in percentage of maintamers and nonmamtam- us selecting general social support

NOIt- General social Mamtarners mamtainers support (n = 63) (n = 41) P value

BOVOW 34.9 42.9 0.30 Take care of house 84. I 83.0 0.80 Adviwopinmn 79.1 19.4 0.99 Confide 65. I 73. I 0.20 Shown concern 85.7 91.5 0.21 Given encouragment 73.0 66 7 0.37 Help manage health 63.5 65.3 0.81 Social activities 93.7 97.9 0.13


tended, overall, to select more family members (e.g. Table 3. Mean specific social support and related variables for

wife, other relatives) than did nonmaintainers. maintainers* and nonmaintaincrst

Another way in which support can be looked at is in the number of individuals identified as supporters. Of the eight questions, only help with managing personal health problems was found to be significantly different (P = 0.05), with maintainers averag- ing 2.2 supporters and nonmaintainers, 1.7. For the other situations, participants in both groups reported a similar pattern in help-seeking and support expecta- tion. For both maintainers and nonmaintainers, one person, on average, was named from whom they would borrow something and in whom they would confide; between one and two people were named who would take care of house, from whom they would seek advice and expect encouragement; and over three people were named who would show concern and with whom they would socialize.

The final two questions dealt with satisfaction with support. No differences between maintainers and nonmaintainers’ were found in satisfaction with the support provided. Overall, participants in both groups indicated a fairly high degree of satisfaction with support; average mean support was 4.07 for maintainers, and 4.01 for nonmaintainers; maximum = 5. Finally, approx. 80% of the respondents in both groups indicated that they were satisfied with the level of help/support experienced in the seven situations and in socializing with others.

Category Mean SD Range P value

Informorionladuice

Maintainers 12.5 5.4 3.25 o.ooo2 Nonmaintainers 9.2 5.3 I .23

Appraisal

Maintainers 12.6 5.8 3.25 0.0008 Nonmaintainers 9.5 5.6 I.25

Emotional supporr Maintainers 17.8 6.4 5.25 0.001 Nonmaintainers 14.2 7.3 2.25

Aoailable

Maintainers 13.7 5.6 3.25 0.01 Nonmaintainers 11.4 6.7 I .25

Sarisfaction

Maintainers 16.0 6.5 4.25 0.003 Nonmaintainers 12.7 7.2 I .25

Con/7icr

Maintainers 4.8 2.1 1.13 0.006 Nonmaintainers 3.8 2.2 1.12

l n = 58; 5 maintainers indicated no support network. tn = 131; IO nonmaintamers indicated no support network. (If all

maintainers and nonmaintainers are included variables remain significant.)

questions related to support for risk reduction are found in Table 3.

To summarize, while no major differences between maintainers and nonmaintainers were found in the general support asked for or expected, some interesting trends were noted. The number of individuals included in participants’ support networks was somewhat limited; participants usually sought help/support from only one to three people; and they were least likely to ask for help and most likely to expect to be shown concern and socialize with others. Never- theless, they were fairly satisfied with the support they received; only a small percentage was interested in changing that amount of support. As for personal health problems, about one-third indicated they usually did not seek any help, and of those that did, maintainers would ask for help from a significantly larger number of network members than would nonmaintainers. Finally, while participants in both groups were family-oriented in naming supporters, maintainers tended to identify wives and other relatives more frequently than nonmaintainers.

The Table shows that for each type of support- information/advice, appraisal, emotional support and availability-the maintainers had significantly more support than did the nonmaintainers. Main- tainers were also significantly more satisfied with their support. Finally, while participants in both groups reported very low levels of conflict with network members, maintainers had significantly more conflict with members of their support network. An intervening variable that could affect the total amount of support provided was the number of people identified in the close support network. It is probable that the larger the network, the more support provided. Therefore, a second series of t-tests controlling for size was calculated for the six variables. After controlling for size, only information/advice remained significantly different between the two groups (P = 0.015).

Support for risk reduction

As to who these supporters were, it was again found that maintainers and nonmaintainers were almost unanimous in selecting wives as their first choice for supporter in the close support network. After wives, maintainers most frequently chose other relatives. Nonmaintainers, while initially selecting other relatives, began to progressively select friends more frequently than did maintainers.

The second questionnaire asked respondents only about those individuals whom they had identified as providers of the most help, assistance or reassurance. For each of the persons named, respondents were asked to specify the degree (not at all = I; slightly = 2; moderately = 3; very = 4; extremely = 5) of help/support the individual provided with risk- reduction efforts, satisfaction with the support provided, and conflict with the individual named. Thus, if a respondent named five people and indicated each was extremely helpful/supportive in a particular area, the support score for that question would be 25. Based on this scoring system the mean responses of maintainers and nonmaintainers to each of the six

Table 4. Primary selection for the five members of the support networks of maintainers and nonmaintainers

Selection of Maintainers Nonmaintainers network of (n = 58) (n = 131) members Relationship (%) (%)

I Wife* 91 87 2 Other relative 64 67 3 Other relative 79 59

Friend I5 34 4 Other relative 64 53

Friend 32 41 5 Other relative 66 43

Friend 29 53

*If married participants only were included, these percentages would be 97% and 95%.

254 PATRICK O’REILLY and H. EMERSON THOMAS

Table 5. The mean and the proportion of specific support for maintiiners and nonmaintainers by relationship

Maintainers Nonmaintainers (n = 58) (n = 131)

Relationship Mean Percentage Mean Percentage

Wife 17.16 29 16.69 34 Other relative 13.11 52 12.52 41 Friend 10.50 16 10.2 22 Neighbor: work associate 10.0 3 9.1 4

Table 5 compares data on mean support and percentage of total support by the provider’s relationship to the participant. Mean support was calculated by summing the amount of support a participant indicated was provided in each of the four risk- related support areas divided by the number of individuals identified in that relationship category (maximum = 20). As seen on the Table, wives provided the largest mean support for both groups. They were followed by other relatives, friends and neighbors/work associates. In contrast, other relatives accounted for the largest percentage of support. (Obviously, wives could only be listed as supporters once while other relatives had the potential of being identified up to five times.) Together, wives and other relatives accounted for over 80% of the risk-related support provided to maintainers and 74% of the support to nonmaintainers. The importance of this finding will be clearer after looking at the close social networks of the participants.

In summary,. highly significiant differences were found in all of the risk-factor-related support variables. with more support being provided to maintainers. Most of these differences could be attributed to differences in the sizes of the networks of the maintainers and nonmaintainers. After controlling for size, only provision of information/advice remained significantly different. Finally, as was found with general support, wives and other family members provided the participants with most of the support for risk-reduction efforts. And, maintainers, in comparison to nonmaintainers, sought/expected support

more often from family members, less often from friends.

Components of the social network

As for the network components of the two groups, Table 6 points out similarities in four areas and differences in three other areas. The close networks of both maintainers and nonmaintainers were similar in gender, geographical distance, duration of the relationship, and frequency of contact. For both groups, network members were approximately evenly divided into male and female; a majority of members were less than 5 min away from the participants (56 and 59%); most members had known the participants over 20 years (80 and 77%); and participants had contact with approximately half of their network members on a daily basis (48 and 49%).

The three areas of difference were in content of the relationship, and size and density of the networks. Proportionately, about the same percentage of the two groups’ networks was composed of wives (23 and 26%) but the networks of the maintainers contained significantly more other relatives (54 vs 42%), and fewer friends (20 vs 27%) than the networks of the nonmaintainers. As noted above, maintainers also identified more individuals in their close network (3.8 vs 3.1) and not surprisingly, with more family members in their network, the maintainers’ network were denser than those of nonmaintainers (0.66 vs 0.51). These data confirm earlier findings to the effect that maintainers had closer, more family-centered networks than nonmaintainers.

Multivariate analysis of the data

The above analyses produced a set of independent but significantly associated variables, and it was considered useful to determine the relationship of these variables to maintenance when analyzed together. Because the analysis required a comparison of two groups, it was decided that the more appropriate approach was the use of the discriminant function. The direct discriminant function analysis enabled us to specify the degree to which these dependent

Table 6. Social network of maintamers and nonmaintzamers in percentages and means

Category

Gender Male Female

Localion (%) 5 min 540 min

60 min Duration (%)

IOyr l&l9 yr 2&29 yr 30 yr

Frequency o/ con,ncr (%) Daily Weekly Monthly

conrent (%) Wife Other relative Friend Neighbor/co-worker

Size (mean) Densitv (mean)

Maintamers Nonmamtainers

48 50 52 50

56 59 3x 35

6 6

6 8 I4 15 31 48 49 29

48 49 I8 14 34 37

23 26 54 42 20 27

3 5 3.8 3.1 0.66 0.51

P value

0.7

0.7

0.89

0.46

0.012

0.019 0.021


Table 7. Number of observations and percentages classified mto each

group

Maintainers Nonmaintainers

Maintamers 42 16

72.41% 27.59%

Nonmamtainers 37 92

28.68% 71.32%

variables correctly predict whether an individual would be a maintainer or a nonmaintainer. The specific variables included in this analysis were: all four risk-related support variables-information/advice, appraisal, emotional support, and availability-plus satisfaction with support, and four network variables-size, density, and the content variable divided into other relatives in network and friends in network.

As can be seen from Table 7, the direct discriminant function analysis with the selected support and network variables successfully classified the maintainers in 72.4% of the cases, and the nonmaintainers in 71.3% of the cases.

The discriminant analysis was also used in a step- wise fashion. This procedure allowed for the de- termination of which of the variables in the analysis were significant, and what amount of variance between the two groups was explained by the significant independent variables.

As seen in Table 8, the three variables that remained significant after the analysis were provision of information/advice on risk-reduction efforts, other relatives in the network, and availability of supportive network members. The canonical correlation is 0.103. This correlation is similar to the correalation ratio and is the total proportion of the variance between the two groups explained by the variables.

DISCUSSION AND CONCLUSIONS

The support questions: what is provided?

The operationalization of social support continues to be a major unresolved issue for many social scientists [21. 37. 381. Lacking a standard for measurement, two distinct instruments were devel- oped for this study to better understand the impact of support on maintenance of improved health status. One questionnaire addressed the issue of general support. that is, support in everyday life, and a second questionnaire was used to look at the impact of support specific to risk reduction. In both ques- tionnaires. support was operationalized as a multi- component variable.

No significant differences were found between the maintainers and the nonmaintainers in the general support they usually asked for or expected to receive. However, the members of both groups did indicate

differing levels of network orientation in the eight situations described. Members of both groups were least likely to ask to borrow something and confide in someone, and on average, would usually ask for help from one person for these types of support. They were more likely to seek advice, expect encouragement and ask for help with health, and named between one and two people for these types of support. The participants were most likely to socialize with others and expect to be shown concern, selecting, on average, over three people for these activities. These differences in seeking or expecting support could be interpreted as a matter of choice and not constraint. That is, the act of asking for support appears to be more of a barrier or limiting factor in the actualization of supportive behaviors from network members. rather than the result of there being no one available to ask.

One of the problem factors related to this reluc- tance to ask for help and the apparent satisfaction of the participants with the support they obtained is the role expectations of these individuals. The male- dominant, decision-making provider role is a very traditional and still acceptable view for the men in the age cohort encompassed by this present study. Any increased sense of dependency, occasioned by the use of instrumental, affective, or other support, may conflict with the way these individuals have envi- sioned their role, particularly their role within the family. Consistent with this role would be a reluc- tance to ask for or expect close family or friends to provide assistance with problems or concerns that are often described as personal.

However, on the positive side, the participants’ willingness to expect and accept the concern of others and to seek advice and information may function as a kind of ‘foundational support’, that is, support that may lead to the acceptance of help for specific health problems. These speculations aside, the principal finding from these data was that support provided in common situations did not differentiate maintainers from nonmaintainers.

In looking at the support specifically provided for risk-reduction efforts, clear differences were found. Maintainers received significantly more support than nonmaintainers in all four areas: information/advice, appraisal, emotional support and availability. How- ever, after controlling for size of the network most, but not all, of the differences between the participants in the two groups was eliminated. The one remaining significant difference was in the amount of information/advice provided. This result shows that for the most part, maintainers and nonmaintainers did not perceive themselves as obtaining differing amounts of support from individuals in their networks. What was different was the maintainers’ perception of the availability of a larger number of supporters. Individ- ual network members were not giving the maintainers

Table 8. Significant variables in step-wse discrtminant analysis

Variables

Information/advice Other relative

Availability

Partial

R?

0.07 0.02

0.02

Average square

Prob. canonical Probabihty F >F correlation > associate

13.99 o.Oc02 0.0703 0.0002 0.076 0.076 0.0862 0.0003

0.065 0.065 0.1031 0.0002

256 PATRICKO'REILLY and H. EMERSON THOMAS

more support than that given nonmaintainers; rather, more individual’s were giving them support.

variables may all have influenced these contrasting results.

In assessing these findings it is important to consider if the different results for general support and support specific to risk reduction may have been due to differences in the way they were measured. The general support questionnaire only asked respondents if support would be asked for/expected in a situation and if so, from whom. The format for risk-reduction support began with a list of supporters about whom respondents were asked to specify the degree of support that each usually provided. Recog- nizing these differences in format, there is a way in which the results from both measures can be reason- ably equated. First, as noted above, risk-reduction support was found to be related largely to the size of participants’ networks; the greater the number of supporters, the more support provided. It was also shown that in the general support areas, with the exception of asking for help with personal health problems, no differences were found between maintainers and nonmaintainers in the number of supporters they identified. Therefore, it is likely that in the general support areas the degree of support provided to maintainers and nonmaintainers is not significantly different. The one area where differences in number of supporters was found, help with personal health, is consistent with differences found between the two groups in the risk-reduction support areas.

A clear implication of these findings, particularly if they are verified by others, is that in order to clarify what relationship social support may have to particular health issues, the operationahzation of support should be specific to the health outcomes of the study. Also, to the extent that the specificity of the instrument is important to its validity in a study, then attempts to develop an instrument that will be gener- ally accepted as a valid measure of social support for most studies may be an unattainable quest rather than a realistic goal.

The network questions: who are the providers?

From the intersecting circles of affiliation [41] that comprise social networks, one may have a number of choices from whom to seek or expect help. This supposition leads to the principal network question addressed in this study; namely, which types of relationships are associated with providing support for maintenance of improved health status?

While somewhat cumbersome, this line of argu- ment appears io provide reasonable evidence to conclude that general instrumental, confidant, appraisal, informational and emotional support was not related to maintaining improvements in cardiovascular health status. This support may, however, be a necessary precondition for the other more specific support for risk-reduction efforts that was found to be very significantly related to maintaining an improved health status.

It was found that participants in both groups selected wives as primary supporter in most of the general support areas and in all of the support areas specific to risk reduction. The participants also perceived wives as providing them with a high level of support in each of the risk-reduction areas of support. Thus, it does appear that from the perspective of the men in both groups, wives not only were the primary person turned to for support but they provided almost as much support as could be asked for or expected. Also, the low level of conflict reported within the network suggests that this level of support was not diminished by any major, ongoing conflict within their relationships.

These results are particularly interesting because Schaefer et al. (271, using the questionnaire from which many of the current support questions were adapted, found no connection between physical health and social support. Other researchers similarly have reported little or no relationship [22, 39,401. The lack of agreement in the findings of this study with past investigations may be due to the specificity of the questions used. Previous studies, including that by Schaefer [27], examined relationships between general support and general and/or specific health issues. The present study sought to show a connection not only between general support and a specific health outcome (i.e. maintenance of improved cardiovascular health status), but between support specific to risk- reduction efforts and the specific health outcome. Similar to other studies, no differences between maintainers and nonmaintainers were found for general support and maintenance, but using the more specific measures of support for risk-reduction efforts, very significant differences were found between them.

For Litman [42], the wife-mother is the central agent of cure and care in the family context. In many studies of help-seeking [43,44], compliance [4547], and rehabilitation [48-511, the support of the wife was of paramount importance. This study reafhrms these earlier findings in pointing out the important role of wives in supporting spouses’ efforts at maintaining an improved health status.

These findings also contrast those reported by Mermelstein et al. [18] who did not find a significant relationship between maintenance of smoking cessation and general support or support specific to smoking cessation. Differences in population, format of instruments. time of administration and independent

However, for three reasons one should not auto- matically conclude that efforts to improve maintenance should focus on increasing spousal support. First, from the perspective of both maintainers and nonmaintainers, wives appear to be already providing a very high level of support for risk-reduction efforts; it is possible that only a limited amount of additional support could easily be drawn from them. Secondly, as Belle (521 has pointed out, an overreliance upon wives for a variety of supportive behaviors may be deleterious to their own health, and in the end the positive function of support (i.e continued, improved health status of the husband) could result in an equally negative outcome (i.e illness of the wife). So, while spousal support should be encouraged, practitioners should be aware of not overburdening a spouse with supportive responsibilities. Finally, for this population, most of whom were married, the support provided by wives did not differentiate maintainers from nonmaintainers. Spousal support would appear to be necessary, but in itself not sufficient, to ensure the maintenance of improved health status. In


order to identify the type of supporters that differentiated the networks of maintainers from nonmaintainers, it is necessary to go further into the support network.

After wives, the next category of supporter is that of other relatives. For both groups, other relatives provided the largest percentage of support, 54% for maintainers and 42% for nonmaintainers. In addition, when comparing the network content of both groups, maintainers’ networks were found to contain significantly more other relatives than nonmaintainers’ networks, while nonmaintainers’ networks contained significarltly more friends. This difference in relational content of networks is important because it was also found that other relatives tended to provide a greater amount of per/person support than friends; with more relatives in their networks, it would be expected that maintainers (even controlling for size) would be provided with more support from their networks. These findings help to address a question that Fischer [29] raised; namely, in the absence of relatives, can friends be substituted? The answer, if support is for maintenance of an improved health status, is probably not. Friends simply do not provide the same level of support for risk-reduction efforts that families do. There are probably exceptions to this statement and instances where friends act as family and may even be identified in terms of familial relationships [53]; nevertheless, it probably is not possible to substitute equally friend support for family support, at least for individuals such as those reported in this study.

Thus, in terms of the participants’ networks, this study found no differences between maintainers and nonmaintainers in the gender of network members nor in their location, duration of relationship or frequency of contact. The data did demonstrate that maintainers had larger, closer and more famly-centered networks than did nonmaintainers. These findings confirm the importance of understanding the social, particularly the familial, context in which changes in personal health behaviors are made and sustained. Successful maintenance, for this population, was not a simple issue of personal responsibility but was clearly related to the degree of supportive response from family members.

Finally, the nine significant variables used in the discriminant function analysis were found to successfully predict maintainers and nonmaintainers in 72% of the cases. This predictive value is modest, but the use of discriminant analysis in support and network research is somewhat unusual, and as such this beginning effort shows promise. The. analysis also points out the limited value of using only social support and social network variables to predict such a relatively complex outcome as maintenance of improved health status. The addition of other variables in future research, drawing on individuals’ personal health and/or medical experiences as well as their psychosocial make-up could expand the predictive power of this approach.

In the indirect discriminant analysis, three of the nine variables were found to significantly discrimi- nate between maintainers and nonmaintainers. These were: information/advice, other relatives in the network, and perceived availability of supporters. These

three variables accounted for 10.7% of the variance between maintainers and nonmaintainers. All of the variables in the analysis tended to be highly corre- lated. so it is not surprising to see only three of the variables significantly related in the step-wise analysis. This analysis would indicate that for this population, the provision of information/advice on risk-reducing efforts, through a network in which there was an appreciable proportion of family members who were available to provide support, accounted for a small but significant amount of the difference between maintainers and nonmaintainers. While the difference accounted for was small, it should be remembered that because heart disease remains the leading cause of death and disability in this country, even a modest increase in the number of individuals who maintain a lowered risk for heart disease could have a major impact on the at-risk population-and their families-as well as the many institutions that share the social and financial bur- dens of this disease.

It should be noted that the results of the multivariate analysis do not support the contention of those investigators [54,55] who have proposed focusing upon the structural aspects of social networks and not on the networks’ supportive functions. If the study had only considered structural network variables we would only have found that large, family- centered networks were important in maintaining improved health status. The inclusion of the support variables and the combined analysis of all significant variables in the study allows for critical additions to these conclusions; namely, that maintenance is directly related to the amount of support specific to risk-reduction efforts, and that the provision of information and advice on risk reduction accounted for most of the variance between the two groups of men studied. Results such as these confirm the importance of measuring both support and network variables. Investigators also need to recognize the multiple nature of these concepts, and include appropriate measures of all of the components of these concepts that would be considered critical to the outcome being studied.

In summary, the principal conclusions to be drawn from the data are that:

-No differences were found between maintainers and nonmaintainers in the general support asked for/expected. However, among the different supportive behaviors, participants from both groups were least likely to borrow something and most likely to be shown concern and socialize with others.

-With the exception of help with health, no differences were found between maintainers and nonmaintainers in the number of individuals from whom they would usually ask for or expect general support. Participants named between one and three persons in the seven situations.

-Significant differences were found between maintainers and nonmaintainers in each of the four risk-reduction support areas, particularly information/advice.

-Most of the significant differences between maintainers and nonmaintainers in risk-reduction

258 PATRICKO'REILLY and H. EMERSON THOMAS

support was found to be related to the large size of the maintainers’ support networks.

--In addition to being larger, the support networks of the maintainers were significantly more dense and family-oriented than the networks of the nonmaintainers.

-For all participants, most of the general support and support for risk-reduction efforts was provided by wives and other relatives.

Finally, it should be stated that since the significant support and network variables accounted for only a small amount of the variance between maintainers and nonmaintainers, it would be inappropriate to over-emphasize the importance of these results. Further research is needed to continue to refine questions related to support networks as well as other personal and social factors that may be related to maintenance of behaviors that improve personal health.

It should be noted that the study design limits the generalizability of these conclusions. Using the original Boston University MRFIT population resulted in a study population of white, middle-aged males most of whom were middle-class and married. The appli- cability of the results to individuals who do not fall within these broad sociodemographic categories would be inappropriate. Another weakness of the study is that the social support and social network data is cross-sectional and retrospective. Longi- tudinal data is clearly more appropriate for this type of study. However, the stability of and lack of conflict in the participants’ networks may indicate that provision of support and the participants’ per- ceptions of that support may not have changed appreciably in the years following the end of the original study.

In addition to these findings, the data from this study imply that efforts to increase our understanding of the impact of the social environment on health status will benefit from clearly specifying the individual components of social support and social networks, and including measures of both in future research. Establishing the distinction between these concepts could be particularly important in terms of identifying in what circumstances or situations behavioral (supportive) or structural (network) inter- ventions would be more effective in altering health status. Also, the reasons for providing specific support may affect the number of supporters and the amount of support they provide. Therefore, it would appear that a single instrument to measure a general level of support may not be a very valid approach to accurately measure support in all circumstances. In- stead, where appropriate, the instruments used to measure support could be designed so that they would utilize a similar format, but the content of each instrument would be specific to the health outcome under study.

The programmatic implications of the current research he in the application of these conclusions to treatment settings. It would seem likely that limiting to the individual patient, or even to the patient and his spouse, treatment or maintenance plans for problems that require long-term life-style or behavioral changes will only result in limited success. As an

alternative, practitioners should consider the value of assessing both the availability of family members to provide support and the support that they currently provide to the at-risk individuals. Drawing family members into the web of support and legitimizing their supportive function will likely enhance the probability of long-term success with risk-reduction efforts. This process of drawing in the family could be carried out by utilizing the wife, the key network supporter, as the conduit to other family members. Simultaneously, health/medical practitioners could decrease patient resistance to the act of asking for help/support by encouraging patients to be more accepting of broad, family-centered support for the modifications they should be attempting. These final methodological and programmatic comments are obviously speculative and are included in the expecta- tion that they may stimulate further research in this area.

Acknowledgemenrs-I would like to acknowledge the assistance of Carol Pickering, Ellen McElroy, Carolyn Pearson. Dana Madden, and Anne Lennerton in the collection of the data, Peter Gaccione in the statistical analysis. and Kirsten Levy for advice and assistance in the preparation of the manuscript.

This study was supported on part by grants from the National Center for Health Services Research (HS05349) and from the National Heart. Lung and Blood Institute (HV33106. HL18318).

REFERENCES

1. Goldman L. and Cook F. The decline in ishemic heart disease mortality rates. Ann. intern. Med. 101, 825, 1984.

2. Schwartz J. L. A critical review and evaluation of smoking control methods. Pub/. HIfh Rep. 34,389, 1969.

3. Sackett D. L. The magnitude of compliance and non- compliance. In Compliance with Therapeutic Regimens (Edited by Sackett D. L. and Haynes R. B.). Johns Hopkins University Press, Baltimore, Md, 1976.

4. West D. W., Grahan S., Swanson M. and Wilkinson G. Five year follow-up of a smoking withdrawal clinic population. Am. J. publ. Hlth 67, 536, 1977.

5. Best J. A. and Bloch M. Compliance in the control of cigarette smoking. In Compliance in Health Care (Edited by Haynes R. B., Taylor D. W. and Sackett D. L.). Johns Hopkins University Press, Baltimore, Md, 1979.

6. Dunbar J. M. and Strunkard A. J. Adherence to medical regimen. In Nufrifion, Lipids, and Coronary Hearr Disease (Edited by Levy R. L., Rifkind B. M., Dennis B. H. and Ernest N. D.). Raven Press, New York, 1979.

7. Brownell K. Public health approaches to obesity and its management. A. Rev. publ. Hlrh 7, 521, 1986.

8. Ockene J. K. Benfari R. R., Nuttall R. L., Hurwitz I. and Ockene I. S. Relationship of psychosocial factors to smoking behavior change in an intervention program. Pren. Med. 11, 13, 1982.

9. Coppotelli H. C. and Orleans C. T. Partner support and other determinants of smoking cessation maintenance among women. J. Consult. clin. Psychol. 53,455, 1985.

10. Mahoney M. J. and Mahoney J. Treatment of obesity: a clinical exploration. In Obesity (Edited by Williams B. T., Martin S. and Foreyt J. P.). Brunner/Mazel, New York, 1976.


11.

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

Brownell K. D., Heckerman C. L., Westlake R. J., Haynes S. C. and Monti P. M. The effect of couples training and partner cooperativeness in the behavioral treatment of obesity. Be-hoc. Res. Ther. 16, 323, 1978. Earp J. L., Ory M. G. and Strogatz D. S. The effects of family involvement and practitioner home visits on the control of hypertension. Am. J. publ. Hllh 72, 1146, 1982. Morisky D. E., DeMuth N. M., Field-Fass M., Green L. W. and Levine D. M. Evaluation of family health education to build social support for long-term control of high blood pressure. Hlrh Educ. Q. 12, 35, 1985. Shenkel R. J., Rogers J. P., Perfetto G. and Levin R. A. Importance of “significant others” in predicting cooper- ation with diabetic regimen. Int. J. Psvchiar. Med. 15, 149, 1985-1986. de Araujo Cl., Van Arsdel P. P. Jr, Holmes T. H. and Dudley D. L. Life change, coping ability, and chronic intrinsic asthma. J. psychosom. Res. 17, 359, 1973. Radius S. M., Becker M. H., Rosenstock I. M., Drach- man R. H., Schbert K. C. and Teets K. C. Factors influencing mothers’ compliance with a medical regimen for asthmatic children. J. Asthma Res. 15, 133, 1978. Levy R. Social support and compliance: a selective review and critique of treatment integrity and outcome measurement. Sot. Sci. Med. 17, 1329, 1983. Mermelstein R., Cohen S., Lichtenstein E., Baer J. S. and Kamarck T. Social support and smoking cessation and maintenance. J. Consult. clin. Psychol. 54, 447, 1986. Multiple Risk Factor Intervention Trial Research Group. Multiple risk factor intervention trial. J. Am. med. Ass. 248, 1465, 1982. Coronary Risk Handbook. American Heart Association (No. EM 620-PE), Dallas, Tex., 1973. Broadhead W. E., Kaplan B. H., James S. A. et al. The epidemiologic evidence for a relationship between social support and health. Am. J. Epidem. 117, 521, 1983. Wallston B. S., Alagna S. W., DeVellis B. M. and DeVellis R. S. Social support and physical health. Hlrh Psvchol. 2, 367, 1983. House J. S. and Kohn R. L. Measures and concepts of social support. In Social Support and Health (Edited by Cohen S. and Syme S. L.). Academic Press, New York, 1985. Orth-Gomer K. and Unden A. L. The measurement of social support in population surveys. Sot. Sci. Med. 24, 83, 1987. O’Reilly, P. Methodological issues in social support and social network research. Sot. Sci. Med. 26, 863, 1988. Henderson S., Duncan-Jones P., Byrne D. G. and Scott R. Measuring social relationships: the interview sched- ule for social interaction. Psychol. Med. 10, 723, 1980. Schaefer C.. Coyne J. C. and Lazarus R. S. The health-related functions of social support. J. behac. Med. 4, 381, 1981. Barrera M. Social support in the adjustment of preg- nant adolescents: assessment issues. In Social Networks and Social Supporr (Edited by Gottleib B. H.). Sage, Beverly Hills. Calif., 1981. Fischer C. S. To Dwell Among Friends: Personal Ner- works in Town and Ciry. University of Chicago Press, Chicago, Ill., 1982. Berkman L. F. and Syme S. L. Social networks, host resistance, and mortality: a nine-year follow-up study of Alameda County residents. Am. J. Epidem. 109, 186, 1979. Di Matte0 M. R. and Hays R. Social support and serious illness. In Social Nerworks and Social Support (Edited by Gottleib B. H.). Sage, Beverly Hills, Calif., 1981. Schultz R. and Rau M. T. Social support through the life course. In Social Support and Health (Edited by

33.

34.

35.

36.

37.

38.

39.

40.

41.

42.

43.

44.

45.

46.

47.

48.

49.

50.

51.

52.

53.

54.

55.

Cohen S. and Syme S. L.). Academic Press. New York. 1985. Hirsch B. J. Natural support systems and coping with major life changes. Am. J. commun. Psychol. 8. 159. 1980. Gallo F. The effects of social support networks on the health of the elderly. Sot. Work Hlfh Care 8, 65. 1982. Branch L. G. and Jette A. M. Elders’ use of informal long-term care assistance. Gerontologist 23, 51. 1983. Israel B. A., Hogue C. C. and Gorton A. Social networks among elderly women: implications for health education practice. Hlth Educ. Q. 10, 173. 1984. Thoits P. Conceptual, methodological, and theoretical problems in studying social support as a buffer against life stress. J. Hlth sot. Behar. 23, 145. 1982. Cohen S. and Syme S. L. (Eds) Social Support and Health. Academic Press, New York, 1985. House J. Work Stress and Social Support. Addison- Wesley Reading, Mass., 1981. Berkman L. F. Assessing the physical health effects of social networks and social support. A. Rev. publ. Hlrh 5, 413, 1984. Simmel G. Conflier and the Web of Group A@liarions (Translated by Wolff K. H. and Bendix R. H.). Free Press, New York, 1956. Litman T. J. The family as a basic unit in health and medical care. Sot. Sci. Med. 8, 49.5, 1974. Geertsen R., Klauber M. R., Rindflesh M., Kane R. L. and Gray R. A re-examination of Suchman’s views on social factors in health care utilization. J. Hlfh sot. Behac. 16, 226, 1975. McKinlay J. B. Social network influences on morbid episodes and the career of help seeking. In The Rele- vance of Social Science for Medicine (Edited bv Eisen- berg L.-and Kleinman A). Reidel, Boston, Mass., 1980. Becker M. H. and Green L. W. A family approach to compliance with medical treatment. A selective review. Inr. J. Hbh Educ. 18, 173, 1975. Sackett D. L. and Haynes R. B. (Eds) Compliance with Therapeutic Regimens. Johns Hopkins University Press, Baltimore, Md, 1976. Gervasio A. H. Family relationships and compliance. In Compliance : The Dilemma of the Chronically Ill (Edited by Gerber K. E. and Nehemkis A. M.). Springer, New York, 1986. Hollib N., Eklund P. and Keenan P. Family conferences as an adjunct to total coronary care. Hearr & Lung 4, 767. 1975. Finlayson A. Social networks as coping resources: lay help and consultation patterns used by women in hus- bands’ post infarction career. Sot. Sci. Med. 10, 97, 1976. Croog, S. H. and Levine S. L. The Heart Patient Recovers. Human Sciences Press, New York, 1977. Mumford E., Schlesinger H. J. and Glass G. V. The effects of psychological intervention on recovery from surgery and heart attacks: an analysis of the literature. Am. J. publ. Hlrh 72, 141, 1982. Belle, D. The stress of caring: women as providers of social support. In Handbook of Stress: Theoretical and Clinical Aspecrs (Edited by Goldberger L. and Brenznitz S.). The Free Press, New York, 1982. Aizenberg R. and Treas J. The family in later life: osvcholonical and demonraohic considerations. In IIandbook of the Psycholog; oj Aging (Edited by Birren J. and Schaie K. W.). 2nd edn. Van Nostrand Reinhold. New York, 1985. ’ Hammer M. “Core” and “extended” social networks in relation to health and illness. Sot. Sci. Med. 17, 405, 1983. Wellman B. Applying network analysis to the study of support. In Social Networks and Social Support (Edited by Gottleib B. H.). Sage, Beverly Hills, Calif.. 1981.

260 PATRICK O’REILLY and H. EMERSON THOMAS

I.

2.

3.

4.

5.

APPENDIX

Principal Components qf General Support Questions Specifjc to Risk Reduction Efforts

When you need a special tool or piece of equipment to finish a job around your house or yard. such as painting or repairing something. do you usually: -Try to borrow it from someone. (From whom?)* -Make a trip to the store. When you go on vacation or out of town for a while, and need someone to take care of your pets, to water plants, or just keep an eye on things. do you usually: -Ask someone to look after things in your home.

(Who?)* --Pay someone to look after things. When you have to make important decisions about your life-for example, about your family or about your jot+ do you usually: -Ask the advice or opinion of others before making

a decision. (Who?)* -Make the decision without asking the advice or

opinion of anyone else. When you are worried or concerned about a personal

matter-for example, a family problem or your health or finances-would you usually: ---Confide in others about your concerns. (In whom?)* -Keep such concerns to yourself. When you go through a particularly difficult time-for example, the death of a friend or relative, a family crisis, or a serious illness--do you usually: -Have others who show an interest or concern for

you. (Who?)* -Not have anyone who shows an interest or concern

for you.

*Respondent could name up to six people. tRespondent could name up to five people.

6. When you carry out different activities at home or at work-for example, making repairs, carrying out a new’ project. working on a hobby-would you usually: -Have others to tell you how you are doing. and to

give you encouragement. (Who?)* -Not have anyone tell you how you are doing and

give you encouragement. 7. When you have a personal health problem that requires

you to take medication, change your diet, lose weight. or stop smoking. do you usually: ----Get help and encouragement from others in manag-

ing these health problems, (From whom?)* -Try to manage these health problems yourself.

8. When you are involced with social activities-for ex-

ample. participating in or watching sports, going to dinner. or to the movies--do you usually: --Do them with others. (Wtth whom?) -Do them by yourself.

Principal Components ?f Social Support Questions Specific to Risk Reduction Efforts

The following questions are about the relationships you have with others you are close tot. For each question rate each person named as follows:

1 = not at all; 2 = slightly; 3 = moderately; 4 = very; 5 = extremely.

I.

2.

3.

4.

5.

6.

How much information or advice with your risk- reduction efforts has this person given you? How useful or helpful was it to have this person tell you how you were doing with your risk-reduction efforts? How much does this person make you feel he/she cares about you and your health? When you needed help with your risk-reduction efforts, was this person available to help you? Are you satisfied with the help this person gave you in your risk-reduction efforts? In the past year, have you had any difficulties or conflicts with this person?

role of support networks in maintenance of improved cardiovascular health status

Documents