Journal of Behavioral Medicine

, Volume 35, Issue 3, pp 305–317

The interplay of dyadic and individual planning of pelvic-floor exercise in prostate-cancer patients following radical prostatectomy

  • Silke Burkert
  • Nina Knoll
  • Aleksandra Luszczynska
  • Oliver Gralla
Article

DOI: 10.1007/s10865-012-9416-2

Cite this article as:
Burkert, S., Knoll, N., Luszczynska, A. et al. J Behav Med (2012) 35: 305. doi:10.1007/s10865-012-9416-2

Abstract

This study broadens the current understanding of the role of planning by focusing on the interplay between individual and dyadic planning (i.e. making plans about the target person’s behaviour together with a partner). Self-report data from N = 141 prostatectomy-patients and their partners were assessed at three times within 1 year post-surgery. Direct and indirect effects of dyadic and individual planning on patients’ pelvic-floor exercise (PFE) were tested. Proposed mediators were social support, social control, and action control. Cross-sectionally, the dyadic planning–PFE relationship was mediated by patients’ received support and partners’ provided social control. Longitudinally, mediators of dyadic planning were partners’ provided social control and support. Effects of individual planning on PFE were mediated by action control at baseline only. Also, at lower levels of individual planning, patients’ dyadic planning was more strongly associated with receipt of social control. Results underscore the importance of social factors in the planning process and its mechanisms in health-behaviour change.

Keywords

Dyadic planning Social exchange processes Self-regulation Health-behaviour change Multiple mediation 

Introduction

Recovery from diseases often requires changes in health-related behaviour. Psychological research usually focuses on individual self-regulatory processes that help to predict health-behaviour change (Schwarzer et al., 2008). Individual action planning in its role as a mediator between goal intentions and behaviour change has been suggested as one such key self-regulatory process in bridging the intention-behaviour gap (Gollwitzer & Sheeran, 2006). While at this point, theory and research have concentrated on individual behaviour-change processes including action planning, the role of the close social environment has rarely been directly addressed (Benyamini et al., 2011; Burkert et al., 2011; Prestwich et al., 2005). This is despite the fact that most adults live in close relationships and health-behaviour change thus happens within a social context. In particular, the regulation of health-related behaviour has been shown to be supported by the close network’s involvement (Berkman et al., 2000). Therefore, the present study directly addresses partners’ involvement in behaviour change by investigating dyadic planning (Burkert et al., 2011). Independent contributions of individual and dyadic planning as well as their interplay are studied in a rehabilitation setting with prostate-cancer patients and their partners who face patients’ functional limitations and ensuing need for the uptake of rehabilitative exercise following radical prostatectomy.

Individual health-behaviour regulation

Planning behaviour change individually refers to generating plans on when, where, and how to implement a new health behaviour. This is also known as implementation intentions (Gollwitzer & Sheeran, 2006). Making plans enhances accessibility of specified opportunities, thereby helps to initiate goal pursuit successfully and accelerates behaviour initiation, when the specified opportunity is encountered (Webb & Sheeran, 2007). A meta-analysis by Gollwitzer and Sheeran (2006) revealed that individual action planning has reliable effects on behaviour change (e.g. small and medium effects on healthy nutrition, and medium to large effects on exercise and testicular self-examination). Independent of which behaviour is planned, planning interventions usually benefit those with high intention to act (Sheeran et al., 2005).

Another individual self-regulatory mechanism, action control, aims at reducing differences between behavioural goals and performance. Derived from the model of self-regulation by Carver and Scheier (2002), action control is conceptualized as a negative feedback loop process: an input function (i.e. current behaviour) is compared with a desired reference value or standard (i.e. behavioural goals). If discrepancies emerge, they are subsequently reduced which constitutes the output function in the model (i.e. behaviour change). Accordingly, action control comprises three facets: first, awareness of standards, which means that behavioural standards must be specified and remembered; second, self-monitoring, where behaviour performance must be monitored and compared to the set standards; third, regulation efforts, which comprise adaptation of behaviour if discrepancies between standards and action are detected (Sniehotta et al., 2006). One facet of action control that may be directly facilitated by planning is awareness of behaviour standards. By forming a plan (e.g. exercising for 10 min, three times a day), individuals set standards for future behaviour. The successful implementation of the planned behaviour is then supported by self-monitoring (e.g. already exercised three times?) and regulatory efforts (e.g. start another exercise unit). Previous studies introduced action control as a mediator between individual planning and change in physical activity (Sniehotta et al., 2005) and as one of mechanisms of a planning intervention to enhance physical activity (Fleig et al., 2011).

Dyadic health-behaviour regulation

Dyadic planning refers to specifying together with a partner when, where, and how the individual target person will implement a new behaviour. In contrast to the conceptualization of collaborative implementation intentions by Prestwich et al. (2005), where both partners jointly planned and enacted the behaviour, the proposed construct of dyadic planning entails that plans are enacted individually, because joint action with the partner might compromise goal achievement. This should be the case, for instance, when the partner breaks the arrangement. Prestwich et al. (2005) could nevertheless show that collaborative planning was more efficacious than individual planning in supporting the enactment of regular breast self-examination. Authors attributed this effect to a facilitated recall of plans, with partners serving as memory aids (Prestwich et al., 2005). However, Prestwich and colleagues have not tested the proposed mechanism in their study. Furthermore, Benyamini et al. (2011) investigated the effects of a planning intervention that instructed women to plan their breast self-examination (BSE) together with their husbands. Compared to women in the control group who planned individually, no intervention effect on BSE was found among women with collaborative planning. Instead, women changed their behaviour when husbands were not involved in their health behaviour prior to intervention and husbands’ knowledge on their wives’ behaviour performance increased. Authors discuss that their study may have served as a trigger for involving the husbands in health-behaviour regulation.

We assume that dyadic planning facilitates health-behaviour change by mechanisms resembling those of individual action planning including enhanced action control (cf. Sniehotta et al., 2005). As in individual planning, dyadic planning should also provide planned standards and regulation guidance (Sniehotta et al., 2005). Additionally, we assume that dyadic planning is strongly associated with other social exchange processes triggered in and provided by the co-planning partner. These additional social exchange processes may include social support and social control.

Social support has been used as an umbrella term for a number of aspects of social relations. Although some authors also include structural aspects of social network composition in the definition of social support, we follow less inclusive definitions and focus on functional elements of social exchanges, such as help, advice, encouragement or comfort, representing instrumental, informational, and emotional support exchanged in times of need (Cohen, 1988; Schwarzer & Knoll, 2007). In this more narrow and functional sense, social support can further be subdivided into a prospective perception that support is available if needed, often termed perceived available support. Retrospective reports of enacted support, on the other hand, may come from different perspectives of the supportive exchange and include mobilized, provided, and received support (Dunkel-Schetter et al., 1992). Although findings on the efficacy of received support to alleviate stress remain inconclusive (e.g. Bolger & Amarel, 2007), partners’ exchange of support in the context of disease has been shown to serve important functions for the recipient, the provider, and their relationship (Knoll et al., 2009a, b; Luszczynska et al., 2007). Previous studies investigating the role of received social support in health-behaviour change have shown that people increased their engagement in health-enhancing behaviours such as diet change or physical activity when those behaviours were supported by their partners (Thrasher et al., 2004). However, intervention studies aiming at enhancement of health behaviour-specific social support provided inconsistent results (Glass, 2000) suggesting additional social mechanisms triggered by the interventions.

Social control refers to strategies that aim at influencing or regulating another person’s behaviour (Lewis & Rook, 1999). Positive control strategies, such as direct persuasion or influencing the enactment of desired behaviour by expression of positive emotions, were shown to be positively related with health-enhancing behaviour (Lewis & Rook, 1999). However, positive control strategies are conceptually interwoven with social support as they contain supportive elements by rewarding desired behaviour. With negative social control attempts, on the other side, network members pressure a target person into a desired health-behaviour change. Negative social control comprises nagging or withdrawing affection and is often inefficient (Lewis & Rook, 1999; Knoll et al., in press) or even counterproductive (Tucker et al., 2006).

Partner involvement in planning should, on the one hand, foster partners’ social support provision by directing partners’ attention to the new behaviour (Dunkel-Schetter & Skokan, 1990). This in turn might facilitate health-behaviour change when target persons are committed to their goals (Brunstein et al., 1996). On the other hand, dyadic planning could trigger partners’ negative control attempts when the target person’s effort invested is not sufficient. However, dyadic planning might also reduce detrimental effects of negative control attempts due to partners’ active involvement and common knowledge of the behaviour-change process. Findings from a dyadic planning intervention study with couples indicated that patients instructed to plan dyadically benefited from the intervention due to indirect effects via social control and action control (Burkert et al., 2011).

To date, potential interaction effects of dyadic and individual planning have not been investigated. Concerning patterns of interplay of dyadic and individual planning, dyadic planning might compensate shortcomings in individual planning. If individuals fail to set behavioural standards on their own, dyadically generated plans might serve this function. However, if both dyadic and individual planning strategies are used, beneficial effects might only be expected if the contents of dyadic and individual plans are similar. Otherwise, distinct behavioural standards might impede self-regulation due to goal conflicts (Emmons & King, 1988).

Pelvic-floor exercise in prostatectomy patients

The present study focuses on the adoption and maintenance of a new health behaviour, pelvic-floor exercise (PFE), recommended to prostate-cancer patients who underwent radical prostatectomy. Prostate cancer is one of the most prevalent cancers in men (Ferlay et al., 2001). In radical prostatectomy (RPE), as one standard treatment for localized prostate cancer, the prostate gland and with it the tumour are surgically removed (Roumeguere et al., 2003). Patients receiving RPE are faced with a number of potential sequelae of the operation. These include erectile dysfunction and urinary incontinence. Incontinence sets in immediately postsurgery and recedes within the first 12 months in most patients (Basillote et al., 2004). Although findings about the efficiency of PFE following RPE are somewhat inconsistent (Hunter et al., 2007; MacDonald et al., 2007), patients are recommended to implement pelvic-floor exercise in order to support recovery from incontinence. There are no guidelines on recommendations of frequency and intensity of pelvic-floor training. However, several exercise units each day should strengthen the outer pelvic-floor muscles. With regard to timing of PFE, a systematic review by MacDonald et al. (2007) revealed that pelvic-floor exercise hastens the return to continence after prostatectomy especially during the first 6 months post-surgery. Patients are nevertheless encouraged to continue exercising at least as long as incontinence persists (Dorey et al., 2009).

Aims and hypotheses

This study investigated unique and interactive effects of dyadic and individual planning on pelvic-floor exercise in prostatectomy patients postsurgery. We assumed that dyadic planning and individual planning both facilitate pelvic-floor exercise by fostering action control (Sniehotta et al., 2005). Additionally, dyadic planning should be associated with other social exchange processes, such as social support and social control (Lewis & Rook, 1999). In order to represent different perspectives on social exchange processes, patients’ and partners’ reports were analysed.

As interactive effects of dyadic and individual PFE-planning have not been tested yet, we tested if deficits in individual PFE-planning would be compensated by dyadic PFE-planning (hypothesis 1). Moreover, we investigated the mechanism through which dyadic planning may operate. In particular, we postulate that associations of dyadic PFE-planning with behaviour would be mediated by social support, social control and action control (hypothesis 2), whereas associations of individual PFE-planning with pelvic-floor exercise should be mediated by action control only (hypothesis 3).

Hypotheses were analysed cross-sectionally involving assessments at 2 weeks postsurgery, and longitudinally with two lags of 6 months each. Measurement points were chosen to represent short-term adaptive processes at 2 weeks postsurgery, when patients started to exercise at home. Longitudinal analyses were applied to investigate whether social and self-regulatory processes may also trigger behaviour in the long run, when incontinence has decreased and pelvic-floor training has started to habituate.

Methods

Procedure

Data were assessed from couples with patients scheduled for prostatectomy at a Department of Urology at a German University Medical Centre. Inclusion criteria were RPE, comprehensive knowledge of the German language, and living in a heterosexual relationship. Of 532 patients scheduled for prostatectomy during data assessment, 87 could not be asked to participate due to logistic reasons, 163 patients did not meet inclusion criteria or did not consent, 38 withdrew consent or did not provide any data, 99 partners did not consent to participation, 4 partners did not provide any data. Data from 141 couples were analyzed. All patients were treated with clinical standard care. Patients received written information about pelvic-floor exercise 1 day postsurgery that recommended pelvic-floor exercise three times a day for at least 10 min per session as long as incontinence persisted. On the third or fourth day after surgery, physiotherapy staff from the Department of Urology introduced patients to pelvic-floor exercise. Data were assessed using questionnaires at three measurement points in time. A first measurement took place 2 weeks postsurgery (t1). Follow-ups were conducted at 6 (t2) and 12 months (t3) postsurgery via mail. The study was approved by the institutional review board.

Patients’ mean age was 63.5 years (SD = 6.1), partners’ mean age was 59.3 years (SD = 7.8). All couples were married or in a stable relationship and had been so for an average of 32.8 years (SD = 12.8). Approximately one third of the patients (30.5 %) and one half of the partners (50.3 %) reported a maximum of 9 or 10 years of schooling, the other two third of patients (63.9 %) and about the other half of partners (42.5 %) had 12 or 13 years of schooling, 18 participants did not respond. A majority of patients were retired (55.3 %), 33.3 % were employed, and 8.5 % were unemployed, whereas 42.6 % of the partners were retired, 40.4 % were employed, and 8.5 % were unemployed. Regarding pelvic-floor exercise before surgery, 68.1 % of the patients had never heard of pelvic-floor exercise before, whereas only 9.2 % reported that they already exercised prior to surgery. Patients’ mean prostate specific antigene (PSA) levels were M = 9.1 (SD = 11.7) pre-surgery, 14.9 % of the patients were classified in tumour stage T1 (not palpable), 53.9 % in T2 (palpable), 27.7 % in T3 (spread through the prostatic capsule), and 2.1 % tumour stage T4 (growing into adjacent organ). In 6 patients tumors had spread to lymph nodes (N1: 4.3 %). None of the patients whose medical data were recorded, had metastases (M0: 98.6 %) at more distant sites of the body. For two patients, medical data were not recorded.
Table 1

Descriptives and intercorrelations of the central variables

 

M

SD

α

2

3

4

5

6

7

8

9

10

1. m int t1

3.79

0.36

.72

.22**

.11

.19*

.26**

.07

−.04

−.03

.07

.01

2. m dap t1

2.50

1.17

.98

 

.43***

.52***

.29***

.07

.24**

.25**

.23**

.26**

3. f dap t1

2.54

1.16

.96

  

.22**

.30***

−.04

.34***

.31***

.08

.26**

4. m iap t1

3.20

0.89

.94

   

.32***

.04

−.01

.02

.02

.01

5. m ac t1

3.38

2.93

.90

    

.06

.08

.14

−.02

.10

6. m ac t2

2.93

0.91

.94

     

.14

.05

.14

.05

7. m sc t1

1.93

0.81

.72

      

.35***

.45***

.27***

8. f sc t1

1.70

0.72

.72

       

.21*

.36***

9. m sc t2

1.74

0.74

.71

        

.47***

10. f sc t2

1.48

0.58

.74

         

11. m ss t1

2.28

0.82

.87

         

12. f ss t1

2.23

0.78

.83

         

13. m ss t2

1.78

0.69

.85

         

14. f ss t2

1.87

0.68

.79

         

15. m pfe t1

143.2

132.7

         

16. m pfe t3

62.3

77.2

         

17. m relsat t1

5.13

0.76

         

18. f relsat t1

4.98

0.82

         

19. m iciq t1

11.84

4.26

         

20. m iciq t3

5.67

4.28

         
 

M

SD

α

11

12

13

14

15

16

17

18

19

20

1. m int t1

3.79

0.36

.72

.01

−.06

.17*

−.06

−.02

−.01

.09

.01

.03

−.04

2. m dap t1

2.50

1.17

.98

.54***

.40***

.40***

.26**

.10

−.03

.24**

.20*

−.09

−.01

3. f dap t1

2.54

1.16

.96

.43***

.61***

.18*

.32***

.15

−.07

.16

.18*

.01

−.05

4. m iap t1

3.20

0.89

.94

.22**

.20*

.20*

.10

.16

.03

.16

.07

−.16

−.16

5. m ac t1

3.38

2.93

.90

.33***

.33***

.07

.31***

.27***

.18*

.28***

.12

−.01

−.19*

6. m ac t2

2.93

0.91

.94

.20*

.07

.46***

.29***

.06

.46***

.01

.01

.28***

.22**

7. m sc t1

1.93

0.81

.72

.55***

.31***

.38***

.30***

−.03

−.04

.09

.10

.11

−.01

8. f sc t1

1.70

0.72

.72

.27***

.34***

.17*

.27***

−.13*

−.03

.08

−.14

−.08

−.03

9. m sc t2

1.74

0.74

.71

.28***

−.01

.42***

.31***

−.08

−.03

.14

.03

.09

−.03

10. f sc t2

1.48

0.58

.74

.31***

.19*

.35***

.37***

−.12

−.09

.10

.03

.07

.01

11. m ss t1

2.28

0.82

.87

 

.44***

.48***

.33***

.24**

.04

.17*

.15

.09

−.07

12. f ss t1

2.23

0.78

.83

  

.27***

.47***

.17*

.12

.12

.08

−.08

−.13

13. m ss t2

1.78

0.69

.85

   

.39***

.05

.25**

.10

.07

.18*

.05

14. f ss t2

1.87

0.68

.79

    

.06

.23**

.15

.03

.24**

.06

15. m pfe t1

143.2

132.7

     

.56***

−.07

.19*

.11

.13

16. m pfe t3

62.3

77.2

      

.05

.06

.18*

.23**

17. m relsat t1

5.13

0.76

       

.34***

.03

−.10

18. f relsat t1

4.98

0.82

        

−.02

−.01

19. m iciq t1

11.84

4.26

         

.37***

20. m iciq t3

5.67

4.28

          

m male, f female, int intention, dap dyadic action planning, iap individual action planning, ac action control, sc social control, ss social support, pfe pelvic-floor exercise, relsat relationship satisfaction, iciq incontinence; decrease of pfe: F(1,140) = 45.43, p < .001; * p < .05; *** p < .01; *** p < .001

Measures

All variables were assessed at all measurement points in time. For hypotheses testing, pelvic-floor exercise at 2 weeks and 12 months postsurgery served as outcomes. Individual and dyadic PFE-planning 2 weeks postsurgery were used as predictors. For cross-sectional analyses, social support, social control, and action control at 2 weeks post-surgery were used as mediators, whereas for longitudinal analyses, their respective 6-month postsurgery assessments were used. Means, standard deviations, internal consistencies, and inter-correlations of respective variables and covariates are listed in Table 1 (for further details on reliability and validity, see Burkert et al., 2011). If not otherwise stated, response options ranged from 1 = “does not apply at all” to 4 = “applies exactly” and mean scores were used.
Table 2

Rotated component matrix of dyadic and individual planning items based on PCA with varimax rotation

 

Component 1

Component 2

I’ve made a detailed plan together with my partner regarding …

  

dap 1 … when to perform pelvic-floor exercise

.93

 

dap 2 … where to perform pelvic-floor exercise

.93

 

dap 3 … how to perform pelvic-floor exercise

.93

 

dap 4 … how often to perform pelvic-floor exercise

.94

 

I’ve made a detailed plan regarding …

  

iap 1 … when to perform pelvic-floor exercise

 

.89

iap 2 … where to perform pelvic-floor exercise

 

.87

iap 3 … how to perform pelvic-floor exercise

 

.88

iap 4 … how often to perform pelvic-floor exercise

 

.89

dap dyadic action planning, iap individual action planning

Patients reported their pelvic-floor exercise (PFE) indicating on how many days during the past week and for how many times a day they had engaged in pelvic-floor exercise. Moreover, they were asked how much time they had usually spent exercising per unit. All three questions were open-ended. Weekly and daily frequencies and average duration per unit were multiplied in order to obtain a measure of weighted duration of pelvic-floor exercise in minutes during the past week.

Patients’ self-reported individual pelvic-floor exercise-planning during the past week was measured with four items using the stem “I have made a detailed plan regarding…” followed by (1) “when”, (2) “where”, (3) “how”, and (4) “how often I perform pelvic-floor exercise”. Items were adapted from Luszczynska and Schwarzer (2003).

Patients’ dyadic pelvic-floor exercise-planning during the past week was assessed using four items with the stem “I have made a detailed plan together with my partner regarding…” followed by (1) “when”, (2) “where”, (3) “how”, and (4) “how often I perform pelvic-floor exercise” (Burkert et al., 2011). Partners’ dyadic pelvic-floor exercise-planning was assessed in a mirrored fashion. Items that followed an identical stem read: (1) “when”, (2) “where”, (3) “how”, and (4) “how often my partner performs pelvic-floor exercise”.

Patients’ reports of action control during the past week were assessed by six items (e.g. “During the last 7 days I have consistently monitored myself whether I exercise frequently enough”; Burkert et al., 2011).

Patients’receivedpelvic-floor exercise-specific social support during the past 7 days was assessed with six items (e.g. “My partner reassured me that pelvic-floor exercise is important”; Burkert et al., 2011). Partners’providedpelvic-floor exercise-specific support was assessed with the same six items, but mirrored (e.g. “I reassured my partner that pelvic-floor exercise is important”).

Patients’ receivedpelvic-floor exercise-specific negative social control during the past week was measured using three items representing negative control strategies (e.g. “My partner pressured me to exercise”; Burkert et al., 2011). Partners’providedpelvic-floor exercise-specific negative control was assessed at 2 weeks and 6 months postsurgery with the same three items, but mirrored (e.g. “I pressured my partner to exercise”).

Patients’ intention to perform pelvic-floor exercise was assessed by three items using the stem “I intend to …” followed by pelvic-floor exercise on three different levels of difficulty (e.g. “perform pelvic-floor exercise regularly”; Burkert et al., 2011).

Patients’ and partners’relationship satisfaction were measured using a single item by Hahlweg (1996) with six response options, ranging from 1 = “very unhappy” to 6 = “very happy”.

Patients’ incontinence was assessed using the German short form of the International Consultation of Incontinence Questionnaire (ICIQ-SF; Karantanis et al., 2004), representing frequency of, amount of, and impairment by urinary incontinence. Responses were summed and ranged from (0) indicating no incontinence to (21) indicating strong incontinence.

Statistical analysis

Path analyses with observed variables were conducted with AMOS 18 (Arbuckle, 2009) to test dyadic and individual PFE-planning as competing predictors of pelvic-floor exercise. In order to investigate cross-sectional and longitudinal predictions from the patients’ and partners’ perspectives, four models were tested. Cross-sectional (t1) Model 1 included patients’ reports of dyadic and individual PFE-planning as well as their interaction as predictors, received social support, social control, and action control as multiple mediators. Pelvic-floor exercise was the dependent variable. Patients’ intention, patients’ incontinence, and patients’ relationship satisfaction were controlled for. Hypotheses were tested with path-models that were applied to two different time frames, accounting for patients’ reports only, or patients’ and partners’ reports: In cross-sectional (t1) Model 2, patients’ reports of dyadic planning, received social support and social control as well as relationship satisfaction were replaced with partners’ reports of dyadic PFE-planning, provided social support, social control as well as relationship satisfaction. In longitudinal Models 3 and 4, baseline predictors remained unchanged. As covariates, intention and relationship satisfaction at baseline, incontinence 12 months post-surgery were now considered and additionally, baseline pelvic-floor exercise was controlled for. Multiple mediators were replaced by respective measurements at 6 months post-surgery. Pelvic-floor exercise at 12 months postsurgery was now the dependent variable. To test the significance of unique indirect effects, Sobel Z tests were conducted (Sobel, 1982). Their effect sizes were estimated by multiplying each predictor-mediator beta by its corresponding mediator-outcome beta.

Moderation analyses followed Aiken and West’s (1991) recommendations. Predictor variables and the constituents of the respective interaction terms were centred around their grand means. To display and test the interaction effects, the Johnson-Neyman technique as outlined by Aiken and West (1991) was applied to identify the range of values of the continuous moderator in which the effect of the continuous independent variable on the outcome was statistically significant.

Univariate and multivariate outliers were treated as suggested by Tabachnick and Fidell (2001). Cases with standardized scores in excess of 3.29 were reduced to one unit above the next highest value. Multivariate outliers were discarded in order to reduce multivariate non-normality indicated by Mardia’s coefficient and its critical ratio (<1.96; Mardia, 1970).

Results

To examine whether the longitudinal subsample was representative of the initial sample, Time 1 responses of participants who provided data until t3 (n = 114 patients, n = 108 partners) were compared with those who did not. No significant differences were found regarding self-regulatory and social exchange processes, health behaviour and sociodemographic variables. However, there was a difference in timing and occurrence of patients’ rehabilitation stay after discharge, Χ2(1) = 5.70, p < .05: among those who dropped out, fewer patients attended rehabilitation 2 weeks postsurgery. Missing data (t1: 23.3 %; t2: 30.5 %, t4: 31.2 %) were estimated by regression imputation (SPSS 19), including all model variables as well as missing mechanisms in the imputation model.

Principle component analysis

As individual and dyadic planning scales have not been assessed simultaneously before, a principal component analysis (PCA) with varimax rotation was conducted to examine the factorial structure of the dyadic and individual PFE-planning items at baseline (see Table 2). Two components were extracted with Eigenvalues >1 (Component 1: 5.57, Component 2: 1.61). The resulting factor loadings corresponded with the theoretical assumptions. The PCA clearly separated the items from patients’ reports of dyadic PFE-planning loading on Component 1 (0.93–0.94) and individual PFE-planning loading on Component 2 (0.87–0.89), with few second loadings on the other factors.

Hypotheses testing

To examine dyadic and individual PFE-planning as competing predictors for pelvic-floor exercise, cross-sectional and longitudinal path analyses with multiple mediators were conducted. Dyadic and individual PFE-planning as well as their interaction were tested as predictors, action control, social control, and social support served as multiple mediators and pelvic-floor exercise was the dependent variable. Intention, incontinence, and relationship satisfaction were routinely entered as covariates in the cross-sectional Models 1 and 2, whereas past behaviour was controlled for in the longitudinal Models 3 and 4.

Cross-sectional Model 1: patients’ reports only

Initially, direct associations of individual and dyadic PFE-planning with patients’ concurrent pelvic-floor training were inspected. Neither dyadic nor individual planning was bivariately associated with pelvic-floor exercise. When tested competingly, individual rather than dyadic PFE-planning had a total effect on pelvic-floor exercise 2 weeks postsurgery, however, only at a 10 %-level of significance. In the full cross-sectional Model 1 at 2 weeks postsurgery (see Fig. 1), patients’ reported dyadic PFE-planning was positively associated with patients’ received social support and received social control from their partners, but not with individual action control. Patients’ individual PFE-planning was positively associated with patients’ action control and negatively associated with patients’ received social control, but unrelated to received social support. Additionally, the interaction of dyadic and individual PFE-planning showed that dyadic PFE-planning was more strongly related to received social control when individual PFE-planning was low (see Fig. 2). Received social support and action control were positively associated with pelvic-floor exercise, whereas received social control and pelvic-floor exercise were unrelated at this point. This model’s Mardia’s coefficient for multivariate non-normality = 1.38 and its critical ratio = 0.52 were satisfactory after exclusion of seven cases. Only after exclusion, individual PFE-planning was significantly associated with action control, whereas the former association of social control and pelvic-floor exercise became non-significant. The indirect effect of dyadic PFE-planning on behaviour via received support (Sobel Z = 1.93, p = .05) had a small effect of .12, accounting for .12 units of the dyadic PFE-planning-behaviour relationship. The indirect effect of individual PFE-planning on behaviour via action control was small (.06), with Sobel Z = 1.57 (p = .12).
Fig. 1

Cross-sectional Model 1 (patient reports only; n = 134) and Model 2 (patient and partner reports; n = 136), controlled for patients’ intention, patients’ incontinence, and patients’ or partners’ relationship satisfaction; partner reports in bold; standardized solution; *p < .05; **p < .01; ***p < .001

Fig. 2

Conditional effect of patients ‘dyadic PFE-planning on patients’ received social control at different levels of patients’ individual PFE-planning two weeks postsurgery

Cross-sectional Model 2: patients’ and partners’ reports

Replacing patients’ reports of dyadic PFE-planning with partners’ reports thereof, again competitive testing yielded a significant total effect of patients’ individual PFE-planning on pelvic-floor exercise, partners’ dyadic PFE-planning remained unrelated. In the full cross-sectional Model 2 (see Fig. 1), partner-reports of dyadic PFE-planning were strongly related to provided social support and social control, whereas patients’ individual PFE-planning was not. Again, patients’ individual rather than partners’ dyadic PFE-planning was related to patients’ action control. No significant interaction effect was found. Partners’ provided social control was negatively related to pelvic-floor exercise, whereas provided social support was unrelated. Further, patients’ action control was positively associated with pelvic-floor exercise 2 weeks postsurgery. This model’s Mardia’s coefficient for multivariate non-normality = 2.68 and its critical ratio = 1.01 were satisfactory after exclusion of five cases. Exclusion led to a non-significant prediction of patients’ action control by partners’ dyadic PFE-planning. The indirect effect of dyadic PFE-planning on behaviour via provided social control (Sobel Z = −2.55, p = .01) was significant with a small effect size of −.10. The indirect effect of individual PFE-planning via action control on behaviour (Sobel Z = 1.90, p = .06) was also small (.07).

Longitudinal Model 3: patients’ reports only

In the longitudinal Model 3, neither patients’ individual nor dyadic PFE-planning 2 weeks postsurgery were related with pelvic-floor exercise 12 months postsurgery. In the full Model 3 (see Fig. 3), patients’ reports of dyadic PFE-planning were positively associated with patients’ received social support and social control 6 months postsurgery. However, they were not associated with later action control. Likewise, patients’ individual PFE-planning was unrelated to any of the tested mediators. No significant interaction was found. Longitudinally, neither patients’ received social support nor patients’ received social control was related to pelvic-floor exercise 12 months postsurgery, whereas action control predicted later pelvic-floor exercise. This model’s Mardia’s coefficient for multivariate non-normality = 5.17 and its critical ratio = 1.75 were satisfactory.
Fig. 3

Longitudinal Model 3 (patient reports only; N = 141) and Model 4 (patient and partner reports; n = 138), controlled for patients’ past pelvic-floor exercise, patients’ intention, and patients’ or partners’ relationship satisfaction at baseline, as well as patients’ incontinence 12 months postsurgery; partner reports in bold; standardized solution; *p < .05; **p < .01; ***p < .001

Longitudinal Model 4: patients’ and partners’ reports

Considering again partners’ reports in Model 4, neither patients’ individual nor partners’ dyadic PFE-planning 2 weeks postsurgery had a total effect on pelvic-floor exercise 12 months postsurgery. In the full Model 4 (see Fig. 3), partners’ dyadic PFE-planning was significantly positively related with partners’ provided social support and provided social control, but not with patients’ action control 6 months postsurgery. Again, patients’ individual PFE-planning remained unrelated to any of the tested mediators. No significant interaction was found. However, in addition to patients’ action control, partners’ reports of social support and social control provision were associated with patients’ pelvic-floor exercise 12 months postsurgery. The small indirect effect (.06) of dyadic PFE-planning on behaviour via provided social support (Sobel Z = 1.96, p = .05) was significant. The indirect effect of dyadic PFE-planning via provided social control on behaviour (Sobel Z = −1.85, p = .07) was also small (−.05). This model’s Mardia’s coefficient for multivariate non-normality = 4.72 and its critical ratio = 1.79 were satisfactory after exclusion of three cases. Exclusion did not change the predictive patterns.

Discussion

This study, investigating the interplay of individual and dyadic planning in their association with patients’ pelvic floor-exercise following radical prostatectomy, aimed at elucidating the potential mechanisms through which dyadic planning may operate. Behaviour-specific action control as well as social support and social control were tested as multiple mediators of the associations between dyadic or individual planning and pelvic-floor exercise. Patient and partner data were used as predictors.

Not in line with hypothesis 1, dyadic PFE-planning did not compensate for deficits in individual PFE-planning with regard to pelvic-floor exercise, action control or social support. Lack of compensation may be explained by ceiling effects in terms of high individual PFE-planning as a preparatory strategy.

Moreover, when patients’ self-regulation in terms of individual PFE-planning was lower, patients were prone to feel controlled by their partners. This was especially true, when they involved their partners in the process of health-behaviour change by planning PFE dyadically. Such a result indicates a social regulatory function of dyadic planning during the implementation phase of a new health-behaviour. The interaction could not be replicated using partners’ reports of dyadic PFE-planning, however. This might be explained by only medium associations between patients’ and partners’ accounts of spousal control of pelvic-floor exercise (ranging between r = .35 and r = .47).

In line with hypothesis 2, among patients in an early stage of behaviour change, dyadic PFE-planning was indirectly related to pelvic-floor exercise via other social exchange mechanisms, including received social support from the partner and provided social control by the partner (see also Burkert et al., 2011).

Comparing the present findings with results of an intervention trial by Burkert et al. (2011), discrepant results emerged. Burkert et al. (2011) showed that a dyadic planning intervention was related with more self-reported dyadic planning which in turn predicted more spousal control. However, for couples participating in the dyadic planning condition spousal control had less detrimental effects on behaviour change than for participants from an individual planning condition in the earlier RCT (Burkert et al., 2011). In the present study, only from the providers’ perspective was spousal control related with lower levels of behaviour outcomes. The inconsistency between the present study and the intervention trial might point to qualitative differences in couples’ spontaneous dyadic planning as opposed to an intervention-induced planning process. Further investigation into couples’ self-initiated planning activities might thus yield more insight into factors that make dyadic planning more or less efficient. Apart from discrepancies between findings in these studies, an explanation for no or even detrimental effects of social control interaction might involve the costs to recipients’ autonomy or self-efficacy (Martire et al. 2011). Moreover, earlier findings point to further moderators in the social control-outcome relationships including relationship satisfaction (Knoll et al., 2011) or centrality of independence (Martire et al., 2011).

Cross-sectional findings further indicated that during an adoption phase of behaviour-change, receipt of support facilitated behaviour change, whereas during later phases of behaviour maintenance partners’ provided support rather than patients’ received support was associated with behaviour. The concept of invisible support might shed further light on this inconsistency in longitudinal models that used patient- or partner-reports of social exchange variables (Bolger & Amarel, 2007). In invisible support provision, a partner’s assistance remains unnoticed by the recipient and therefore does not affect his perceived autonomy and self-esteem. Findings from cross-sectional models involving partners’ accounts of social exchange factors, however, were not in line with invisible support assumptions. Partners’ reports of provided social support failed to mediate the dyadic planning-PFE relationship cross-sectionally, while patients’ accounts of received support did. An alternative explanation for these results might involve time, practice and salience of support for pelvic-floor exercise. While implementing a new behaviour, such as pelvic-floor exercise, patients might have detected receipt of support for said behaviour more easily. After a few months, however, receipt of support for maintenance of this behaviour might not be as salient and more often go unnoticed. Also initially, patients might have regarded a number of actions of their partners as more helpful than partners thought they had been. Partners, however, were gaining expertise at providing support for pelvic-floor training, possibly shaping their supportive action along the needs of patients, creating a better fit over time. This might explain why reports of received support from patients were more closely associated with patients’ behaviour early on while reports of provided support from partners championed prediction of patients’ behaviour in the months to come.

With regard to action control as an assumed mediator of dyadic planning, dyadic PFE-planning was unrelated to action control in the present study. An explanation could involve conflicting plans. If planning involves setting clear standards in terms of when, where, and how to perform the behaviour, then dyadically and individually formed plans might be conflicting. Conflicting or deficient standards have been shown to lead to underregulation (Baumeister & Heatherton, 1996; Emmons & King, 1988). Patients might have primarily chosen their individually formed plans in order to regulate their behaviour. To support this explanation, future research should investigate differences and similarities in contents of dyadically and individually generated plans. In line with hypothesis 3, individual PFE-planning indirectly affected pelvic-floor exercise via action control, but only cross-sectionally at 2 weeks postsurgery. While action control remained an important unique predictor of behaviour maintenance, longitudinally, neither dyadic nor individual PFE-planning was indirectly related with pelvic-floor exercise via action control. In contrast, findings from longitudinal models indicated that dyadic rather than individual planning was associated with the maintenance of health behaviour via social-exchange mechanisms as reported by partners. These results contradict findings from a dyadic planning intervention study (Burkert et al., 2011), in which action control turned out to be a long-term mediator of dyadic planning.

Limitations and outlook

This study has several limitations. First, dyadic and individual planning were measured with standardized self-report measures and quantitative assessments. Future studies might investigate plan content using individualised qualitative assessments in order to test for conflicting standards. Also, important features of the dyadic planning process itself should be elucidated by using observational assessment techniques.

Because pelvic-floor exercise is strongly associated with patients’ experience of incontinence symptoms, the weekly duration of pelvic-floor exercise decreases as patients recover from incontinence within the first year postsurgery. Moreover, inter-individual differences in patients’ and partners’ coping skills as well as expectations that incontinence can be controlled by means of pelvic-floor exercise, have not been controlled for in the present study but might have influenced the effects of dyadic planning and its social mechanisms on behaviour. Therefore, the tested effects might have been underestimated in the present study. Future studies should investigate other health behaviours as outcomes and additional moderators of the respective effects.

The operationalisation of social support in the present study was limited to enacted forms of pelvic-floor training-specific support, i.e. retrospectively reported patient-received and partner-provided support. Thus, we were unable to address perceived available support as a prominent indicator of expected supportive exchange. Also, possible moderating effects of support mobilisation by patients were not investigated in the present study. Future research might address these shortcomings and investigate the role of perceived support as an alternative mediator and also test if received support that was explicitly requested might be more strongly associated with behaviour change than non-mobilized support.

Moreover, dyadic planning was operationalised in couples of male target persons and female partners only. As social support literature suggests gender differences in social exchange processes, role and gender effects might be confounded in the present results. Future work should test dyadic planning and its mechanisms in mixed samples in order to prevent such confounds. Finally, effects of dyadic and individual planning might have been underestimated due to patients’ high levels of individual planning. Therefore, future research should investigate planning processes in less motivated, non-clinical samples.

Conclusions

The present findings extend the conceptual framework of planning by differentiating between individual and dyadic planning processes, and elucidating potential social and self-regulatory mechanisms by which dyadic planning may operate. Moreover, the present study shows effects of patients’ and partners’ dyadic planning as well as social support and control and thus points to the importance of investigating social exchange mechanisms in addition to self-regulatory mechanisms in the process of health-behaviour change.

Following from this, health-care professionals might point out the importance of patients’ self-regulatory skills for the implementation of regular pelvic-floor exercise, but also actively involve partners into the postsurgical adaptation process as well as encourage couples’ open communication about incontinence and means to relieve it.

Copyright information

© Springer Science+Business Media, LLC 2012

Authors and Affiliations

  • Silke Burkert
    • 1
  • Nina Knoll
    • 1
    • 2
  • Aleksandra Luszczynska
    • 3
    • 4
  • Oliver Gralla
    • 5
  1. 1.Institute of Medical PsychologyCharité, Universitätsmedizin BerlinBerlinGermany
  2. 2.Freie Universität BerlinBerlinGermany
  3. 3.Warsaw School of Social Sciences and HumanitiesWarsawPoland
  4. 4.Trauma, Health, and Hazards CenterUniversity of ColoradoColorado SpringsUSA
  5. 5.Urologie am RingCologneGermany

Personalised recommendations