An evaluation of time-limited psychodynamic psychotherapy for couples: A pilot study

292

Psychology and Psychotherapy: Theory, Research and Practice (2012), 85, 292–309C© 2011 The British Psychological Society

TheBritishPsychologicalSociety

www.wileyonlinelibrary.com

An evaluation of time-limited psychodynamicpsychotherapy for couples: A pilot study

Andrew Balfour1, and Monica Lanman2*1Tavistock Centre for Couple Relationships, UK2Tavistock and Portman NHS Foundation Trust, UK

Objectives. Psychodynamic Couple Psychotherapy has developed as a modality inonly a few organizations in the public and voluntary sectors in this country. Varieties ofcouple therapy have evolved due to economic or other constraints, some more open-ended, others involving differing time limits or behavioural techniques. In this study,a time limit of 40 sessions was imposed on the Psychodynamic therapy to improvecomparability with other therapeutic approaches. We examined work with 18 couples,employing various measures which, while not in the context of a full controlled trial,produced some interesting and indicative results. We aimed to investigate (1) the effectsof time-limited psychodynamic couple psychotherapy, and (2) whether the measuresused produce interesting results after 40 weeks.

Design. Within a normal clinical setting, measurements of individual and couplefunctioning would be taken at fixed points in the course of 40-week couple therapies,and analysed for evidence of significant change. Due to funding and clinical limitationswithin the setting, a baseline period before therapy started was used instead of a controlgroup.

Method. Couples were invited to opt in to the study when applying to the agencyfor therapy. They were provided with 40 weekly sessions of couple therapy. Videotapesof sessions at beginning, middle, and end of the therapies were rated by independentobserver, using the Personal Relatedness Profile (PRP) (Hobson, Patrick, & Valentine,1998) adapted for couples (Lanman, Grier, & Evans, 2003), alongside two individual self-report measures, Clinical Outcomes in Routine Evaluation (CORE) (Evans et al., 2000),and the Golombok Rust Inventory of Marital Satisfaction (GRIMS) (Rust, Bennun, Crow,& Golumbok, 1990).

Results. The couples showed improvement as rated both by therapists and observers(rating the videotaped sessions) on the PRP after 40 sessions. On the CORE measure,participants showed improvement at both 20 and 40 sessions. On the GRIMS measureof marital satisfaction, results were non-significant.

∗Correspondence should be addressed to: M. Lanman, Adult Department, Tavistock Clinic, 120 Belsize Lane, London NW35BA, UK (e-mail: [email protected]).Both authors worked for the Tavistock Centre for Couple Relationships when the study was undertaken.

DOI:10.1111/j.2044-8341.2011.02030.x

Couple therapy 293

Conclusions. The results provide indicative evidence for the effectiveness of 40-session couple psychotherapy and provide some convergent validation for the utility ofthe PRP (as adapted for couples) as a measure of change.

Couple therapy is a valued, but scarce resource in adult and child mental health.It is now increasingly accepted that the adult couple relationship is a key locusof intervention in adult mental health problems, with the potential also to improvethe mental health and longer term outcomes for the children of couples in conflict.Historically, couple therapy has struggled to have its status as a psychological therapyrecognized, although the evidence base for the effectiveness of couple therapies ofdifferent orientations in improving mental health is becoming established, see Dunnand Schwebel (1995) and Baucom, Shohan, Mueser, and Stickle (1998) for meta-analytic reviews. A government policy review document used to drive commissioningof services to couples, Moving Forward Together (2002), concluded pessimistically thatpsychodynamic couple interventions were generally ‘too little, too late’ and did not helpcouples to maintain their relationships, but rather to separate more thoughtfully. Whilstthis may be a possible outcome of couple therapy, this conclusion by no means reflectsthe current evidence base for its effectiveness as a psychological therapy. Difficultiesin establishing intimacy and problems in managing conflict are predictive of depression(Christian, O’Leary, & Vivian, 1994) and studies indicate that couple therapy is effective intreating depression (Bodenmann et al., 2008; Leff et al.,2000). Other studies indicate thatcouple therapy reduces violence in couple relationships (O’Farrell, Murphy, Stephan,Fals-Stewart, & Murphy, 2004). In addition, Simpson, Atkins, Gattis, and Christensen(2008) found that couple therapy works on low-level aggression even without targetingit. Another area where couple therapy has been found to show beneficial impact isin relation to alcohol and drug abuse (Fals-Stewart, Klostermann, Yates, O’Farrell, &Birchler, 2005). Studies have looked not only at the effects of the treatment on thesubstance abusing patient, but also at the impact of the treatment on the functioningof their children (Kelley & Fals-Stewart, 2002). Other studies have compared psycho-educational and therapeutic interventions for couples, and the latter has been found tohave a superior effect not only on the parental relationship but on children’s functioning,tracked longitudinally (Pape Cowan & Cowan, 2005). The importance of couple therapyis underscored by the link to child mental health, with research evidence that shows theimpact of unresolved parental conflict upon the mental health, cognitive, and emotionaldevelopment of children (Cummings & Davies, 1994; Fincham, Grych, & Osborne,1994; Harold & Conger, 1997). The importance of conflict in the parents’ relationshipon the mental health of the child is recognized in National Institute for Health andClinical Excellence (NICE) 2005 guidelines on Child Depression, ‘Family risk factors fordepression in children and adolescents include parent-child conflict, parental discord,divorce and separation’ (p 62–63).

Although often referred to generically as ‘couple therapy’, there is a range of differentorientations within this modality, and as elsewhere in the field of psychotherapy, therelative effectiveness of different orientations is contested. ‘The Guideline for depression’review document from the National Institute for Health and Clinical Excellence (2009)concludes by saying that couple focused therapy for depression should normally bebased on behavioural principles. But this conclusion is curious given the current evidencewhich provides a much more mixed picture. Baucom et al. (1998) compared mean effectsize (ES) over a series of 17 independently controlled outcome studies, and found that

294 Andrew Balfour and Monica Lanman

cognitive behavioural therapy (CBT) for couples was indeed an effective and specificintervention for marital distress and some adult mental health problems, with a highprobability that improvement will be maintained for a year after the therapy. Christensen,Atkins, Yi, Baucom, and George (2006) compared traditional behavioural couple therapywith an integrative behavioural approach (which, in addition to behavioural techniques,includes a focus on increasing emotional acceptance between partners). They foundsimilar levels of clinically significant improvement for couples across the two treatments.Couples in the integrative treatment made steady improvement, as compared with thosehaving the traditional behavioural approach. The latter improved more quickly earlyon, but plateaued later in treatment. Crowe (1978) compared a behavioural approachto couple therapy with an interpretative (psychodynamic) approach and a non-specificsupportive therapy. The behavioural treatment was superior to the supportive therapy,with the interpretative treatment intermediate, and the improvements were maintainedat 18 months follow-up. But the randomized controlled study carried out by Leff et al.(2000) compared systemic couple therapy with individual cognitive behavioural and drugtreatments, for depression. The systemic couple therapy was found to be more effectivethan pharmacological treatment, and was at least as cost effective as medication. Notably,the study had a CBT condition which had such high attrition of participants that it wasdiscontinued. Snyder and Wills (1989) compared the outcome of behavioural versusinsight orientated (psychodynamic) couple therapy, and found that there was equalimprovement in the two conditions, with both being superior to waiting-list controls. Asimilar result was obtained by Emmelkamp, Van Der Helm, Macgillavry, and Van Zanten(1984), comparing behavioural with systemic couple therapy, both producing equalimprovements, and both being superior to waiting-list controls. More importantly, whenthese studies were followed up, the results showed that insight oriented, psychodynamictherapies achieved longer lasting change, even when they were evaluated behaviourally(see Snyder, Wills, & Grady-Fletcher, 1991).

Overall, there is a mixed picture emerging from these studies, with some evidencefor the effectiveness of couple psychotherapy from each of the major therapeuticorientations, cognitive behavioural, systemic, and psychodynamic. One important factorin terms of using research to guide clinical delivery, is the question of how much‘external validity’ these studies have – and how well they reflect the complexity of theclinical situation. Whilst well-controlled Randomized Controlled Trials (RCT’s) remainthe ‘gold standard’ of outcome research, they are not without their limitations. Hotopf,Churchill, and Lewis (1999) point out that patients, clinicians, and decision-makers needto know how treatment works in the real world, and to what extent it is effectiveunder routine conditions. Similarly, Pinsof, Wynne, and Hambright (1996) point out thatthere is evidence to suggest that couple and family therapies are effective, but there areconceptual and methodological problems in the current evidence base. In particular,they point out that this research is composed primarily of efficacy studies and the extentto which it pertains to the real clinical situation on the ground, with co-morbidity andcomplex problems, is unclear. These authors point out the urgent need for researchersto address this, which we do, in so far as ours is a naturalistic study of couple therapy asdelivered in a clinical setting.

Psychodynamic couple psychotherapy has been researched and evaluated in fewerstudies than other modalities (Boddington and Lavender, 1995). As Richardson (2003)pointed out, a lack of evidence does not amount to evidence for lack of effectiveness.Researchers seeking to evaluate the psychodynamic approach face particular challenges,in that such interventions are designed to help couples explore how difficulties have

Couple therapy 295

arisen in their relationship and what interferes with changing them, looking at whatlies behind current difficulties and paying attention both to conscious and unconsciousfactors. There is a need for measures that have specificity to the couple modality andsensitivity to the changes that it aims to effect in couples’ functioning. Our studyaddresses this, using a measure of individual psychodynamic functioning adapted forcouples, the Personal Relatedness Profile (PRP) (Hobson, Patrick, & Valentine, 1998;Lanman, Grier, & Evans, 2003), and exploring its utility as an outcome measure that canpick up the changes in qualities of relatedness we are looking for.

Outcome will be influenced by how much time is allowed for the therapy, andin the follow-up. Current NICE recommendations (2009) suggest that an adequatecourse of therapy should be 15–20 sessions over 5–6 months, while practitioners ofa psychodynamic approach often contend that such change takes longer.

This pilot study had two objectives: (1) to evaluate the effect of psychodynamicpsychotherapy for couples delivered over 40 weekly sessions in a naturalistic setting; (2)to explore the utility of a psychodynamic instrument (the PRP as adapted for couples,Lanman et al., 2003) as a measure of change in psychodynamic couple psychotherapy.Our hypotheses were that, after 40 sessions of psychodynamic couple psychotherapy:(1) Couples would show improvements in patterns of interpersonal relating assessed byobserver and therapist ratings on the PRP; (2) they would also show improvements inpsychological state, as measured on individual self-report (Clinical Outcomes in RoutineEvaluation [CORE], Evans et al., 2000); and (3) on self-reported relationship satisfaction(Golombok Rust Inventory of Marital Satisfaction [GRIMS], Rust, Bennun, Crow, &Golumbok, 1990). In addition, we tested whether there would be changes on thesesame self-report measures between a point 3 weeks prior to intake and the start of thefirst session (a ‘baseline’ period).

MethodThe participantsThe 36 individuals making up our 18 couples ranged in age from 28 to 53. Their meanage was 38.35 years (SD 6.45), the mean age of the men being 39.67 (SD 6.61), and ofthe women 36.88 years (SD 6.13).

Four of the 18 couples had three children, four had two children, five had one child,and five had none, although one of those last five couples fostered babies short-term. So,most of our couples were in the midst of childrearing, three with small babies, and justone with more or less grown up children. Ethnically, the couples were predominantlywhite English or European, but two individuals were from Latin America, one was blackCaribbean, and two were Indian. The study being set within a voluntary sector, fee-charging service located in north-west London, meant that those who volunteered totake part expected to pay something for their therapy, although they were charged theagency’s minimum rate. All but one had at least one partner in employment. In one case,both partners worked part-time; in six couples the female partner worked full time, twowives worked part-time, and the remaining 10 women were not in formal employment.

In terms of indicators of co-morbidity, seven couples came to the study with onepartner complaining of depression (three were on medication). Clinical judgementindicated the following: two patients had personality disorders; three had histories ofdrug or alcohol abuse, and one of bulimia. In 12 of the couples, one or both partners hadtried some other form of counselling or psychotherapy previously. The CORE therapists’


assessments indicated levels of anxiety and or depression causing moderate or severedifficulty in 27 participants, personality difficulties in 13, ‘mild’ risk of suicide in three,and of harm to others in three.

Couples complained of a variety of difficulties when they entered the study. Oftenthese were to do with one partner feeling shut out, or over-controlled by the other,or seen as too passive, dependent, or needy. Frequently they complained of rowingconstantly, often over money, sex, and the difference the arrival of children had made.Sometimes the rows included fear of, or some actual, violence, and infidelity was cited bythree couples. Several were on the brink of separating. Those relationships recovered,although for two couples in the study, separation seemed, in the end, to be the bestoutcome.

The therapistsFive therapists were used, all of whom were trained to a very high level in the TavistockCentre for Couple Relationships’ model of psychodynamic couple therapy. This is a4-year training, or a 2-year one on top of previous individual psychotherapy training,and includes personal psychotherapy at least three times per week for the duration ofthe training. While for this pilot, resources were not available to develop a manual forthe therapy and systematically monitor adherence to it, there was regular monitoringthrough case discussion with colleagues, based on video tapes of couples in therapy.Preparation for the study consisted of careful communication of the procedures involved,and peer group discussion of the modification of technique required for time-limitedtherapy. Modification was minimal in that the therapeutic work was conducted ‘asusual’, involving the study of the interactions unfolding in the consulting room, withan additional ‘reality’ of limited time impinging. For a discussion of this technicalmodification see Lanman and Grier (2001).

MeasuresPersonal Relatedness Profile (PRP), (Hobson et al., 1998), adapted for couples, (Lanmanet al., 2003).

The PRP is a 30-item measure which operationalizes fundamental psychoanalyticclinical concepts, assessing the qualities of a person’s ‘internal object relations’ basedon the notion that there are patterns of interpersonal relatedness which may becharacterized according to whether they are ‘paranoid-schizoid’ or ‘depressive’ in quality(Segal, 1973). Psychodynamic therapy aims to effect unconscious changes, such thatpatterns of interpersonal relatedness will be less ‘paranoid’, a state of mind characterizedby projection, distortion, lack of separateness, feeling persecuted, and persecuting inrelation to the other. Instead, relating becomes more ‘depressive’, that is, characterizedby greater reflexiveness and capacity to take responsibility for oneself and one’s ownfeelings, less likely to see the blame as lying only outside the self.

In terms of the validity of the measure, Hobson et al. (1998) began to establish this,initially by testing whether the two axes of the scale (‘depressive’ and ‘paranoid schizoid’)would differentiate between groups of borderline and dysthymic patients. Althoughmore work needs to be done to establish the validity of the measure, Hobson et al.present evidence that their attempt to operationalize paranoid schizoid and depressivefunctioning does tap a meaningful constellation of clinically grounded phenomena.

The couple adaptation of the PRP (Lanman et al., 2003) involved amending theinstructions to raters, in line with dynamic couple theory, so that they considered the

Couple therapy 297

couple as one. This meant observing couples in terms of a shared mode of relating, anapproach with a strong conceptual basis in clinical practice and theory (Ruszczynski,1993). The two people who form a couple are considered to share a predominantmode of psychological functioning (of which they may not be conscious), such thata psychodynamically trained observer can watch them interacting and make a reliablejudgement about the nature and quality of that functioning, ‘as if they were watching oneperson’. A high degree of reliability was obtained when videotapes of couple sessionswere rated independently by seven raters (Lanman et al., 2003). The strong inter-raterreliability found for the original measure by Hobson et al. (1998) (ICC = 0.91) wasreplicated by Lanman et al. (2003), using the couple version of the measure, who foundoverall inter-rater reliability using seven raters independently rating the interactions onvideotapes of 19 couples (ICC = 0.92). In the current study, similar levels of inter-raterreliability were found (Paranoid Schizoid [PS] scale ICC = 0.93; Depressive [DP] scaleICC = 0.91). The PRP has not been used as an outcome measure before this study.

Clinical Outcomes in Routine Evaluation (CORE) (Evans et al., 2000) is a measureof individual mental state. This measure is gaining ground in the United Kingdom as astandardized measure of outcome. It is an individual patient self-report questionnairecomprising 34 questions about how the patient has been feeling over the last week. Itcovers four dimensions: subjective well-being, problems/symptoms, life-functioning andrisk/harm, yielding a mean score to indicate the level of current global psychologicaldistress. The aim of the measure is to provide a pan-theoretical battery to assess changemeasuring the ‘core’ domains of problems. It is designed to be repeated at intervals, withpre- and post-therapy scores used as a measure of ‘outcome’, and has been validatedwith samples from the general population (Connell et al., 2007), NHS primary (Mellor-Clark, Connell, Barkham, & Cummins, 2001) and secondary (Barkham, Gilbert, Connell,Marshall, & Twigg, 2005b) care, and with older adults (Barkham, Culverwell, Spindler,Twigg, & Connell, 2005a). Internal and test–retest reliability are good (ICC = 0.75–0.95),as is convergent validity with seven other instruments and large differences betweenclinical and non-clinical samples; it has also been shown to have good sensitivity tochange (Evans, Connell, Margison, & Mellor-Clark, 2002).

The instrument yields a total score, which is a composite for each individual of theirscores on the individual domains of well-being, symptoms and functioning, not includingthe ‘risk’ sub-scale. This approach, recommended within the CORE system handbook(Barkham et al., 1998; Evans et al., 2002), is the one used here. In addition, the COREoutcome measure provides norms which allow us to judge whether a CORE score isreflective of a ‘clinical population’ or not. This clinical/non-clinical categorization wasestablished by comparing the scores of a large normal population sample with the scoresof a large sample of patients in therapy (Barkham, Mellor-Clark, Connell, & Cahill, 2006).It provides two measures of change: significant change, that is, exceeding that whichwould be expected by chance alone, and ‘clinically significant’ change, which is indicatedwhen a patient’s CORE score moves from the clinical to the non-clinical population. Thecut-off scores for the latter that we have used to indicate clinical population are thefollowing: for males, scores above 1.29 and for females, 1.19 (Evans, 2003).

Golombok Rust Inventory of Marital State (GRIMS) (Rust et al., 1990) was selectedbecause it is a British measure of couple functioning, and so more culturally attuned thanmeasures originating in the United States. GRIMS is a 28-item questionnaire designedto be completed by both partners individually. The authors report split-half reliabilityof 0.87 for women and 0.91 for men, as well as good content and discriminativevalidity. Rust et al. (1990) use norm-referencing and criterion referencing to indicate


the meaning of scores in relation to the severity of relationship difficulty (higher scoresindicate poorer marital quality, and scores in the range of 30–33 represent averagemarital satisfaction). The items also have good face validity, and the authors emphasizethe diagnostic validity, and conceptual grounding that the scale has, drawing its itemsfrom widespread consultation with couple counsellors and therapists. Their intentionis that the scale taps issues ‘known to be important to a good relationship and whichmay therefore be expected to show change when a couple is treated for relationshipdifficulties’.

It is important to bear in mind that the measures look at very different things.Both CORE and GRIMS are self-report measures designed to pick up feelings whichthe individual is consciously aware of. The CORE looks at their individual levelof unhappiness or well-being. The GRIMS looks at the couple’s conscious level ofsatisfaction with the relationship. The problem is that a couple, or one or other ofthe partners in it, might become more concerned about the relationship, or even moretroubled, as a result of becoming more aware of what is going on in the relationshipand why. This could well be understood as being a good outcome of psychotherapy,certainly a first and often necessary step towards improving things in the longer term,although it might not show up in a questionnaire about ‘satisfaction’. The adapted PRPis different from the other two measures in that it is designed to pick up a couple’s waysof functioning of which both partners may be unaware.

ProcedureAll couples applying or being referred to the Tavistock Centre for Couple Relationshipsin the designated period were invited to participate in the study. The offer involvedtheir having 40-session, rather than open-ended therapy, agreeing to their sessions beingvideotaped, and completing questionnaires. Some couples in the study were referredafterwards to the open-ended service. The conditions of the study were close to routine.The therapies were conducted in an ordinary clinic consulting room, albeit one equippedwith a discreetly placed camera. Participants were treated by trained couple therapistsas part of their regular clinical caseload. The observer-raters for the PRP were similarlytrained, a requirement discussed by Lanman et al. (2003). The use of the adapted PRPdepends on an understanding of couple functioning at a level which may lie outside theconscious awareness of the couple, but therapists trained in a psychodynamic approachto couple therapy can reach a sufficient level of agreement when independently ratingcouple sessions using the instrument.

Of the 30 couples who expressed interest in participating, nine withdrew from theresearch during the assessment process and, with the therapists’ support, were referredelsewhere, some into the open-ended service because they did not want the 40-sessionlimit, or decided they did not want to complete questionnaires and be video-taped.Only one couple was turned down by project staff as needing some form of help otherthan psychotherapy. Three couples started participating but withdrew from the therapybefore completing 40 sessions. Finally 18 couples completed the full 40 sessions, and itis these couples which constitute our participant group.

Videotaped sessions taken from the beginning, middle, and end of therapy wereobserver-rated using the PRP, and at the same time points patients completed thetwo individual self-report measures of outcome, the CORE and the GRIMS. In addition,therapists themselves were asked to rate couples using the PRP. We decided that ‘endof therapy’ data would be taken from the penultimate rather than the last session, in line

Couple therapy 299

with the clinical view that the latter is likely to be distorted by either a resurgence ofpresenting symptoms, or a wish to ‘tidy up’ as a response to ending.

To maximize ‘blindness’ in the observer rating, stratified randomization of thevideotaped sessions was used, with the constraint that we ensured that two sessions fromthe same couple were never rated contiguously or in the correct temporal sequence.

Baseline periodWe assessed the stability of the self-report measures of the GRIMS and CORE by askingparticipants to complete them at 3 weeks prior to therapy and then at initial consultation.

Independent observer ratings using the PRPThe PRP consists of a single scale, but two principal axes can be isolated for the purposeof the analysis, the DP and the PS. The reliability of the independent ratings usingthe couple-adapted PRP was assessed by looking at the inter-rater agreement betweentwo raters on both the PS and DP scales for six sessions drawn from six differentcouples at varying time points. Inter-rater reliability was found to be high for the PSScale (Intraclass Correlation = .93, 95% confidence intervals .79–.99) and also for the DPscale (ICC = .91, 95% confidence intervals .78–.98). This is in line with previous studiesof this measure that indicate high inter-rater reliability for the PRP (Hobson et al., 1998;Lanman et al., 2003). The high level of agreement obtained between two raters on thissub-sample allowed us to use one rater for the PRP assessment of the couples in ourstudy.

Repeated measures analyses of variance (ANOVA) were carried out on the PRP, theCORE, and the GRIMS data. The CORE data were also analysed for clinically significantchange using Chi-Square. The data were analysed using SPSS for Windows version 17.

ResultsHypothesis (1) that ‘Patterns of interpersonal relating’ between partners in the coupleswould show improvement, that is, higher scores on the ‘Depressive’ axis and lowerscores on the ‘Paranoid Schizoid’ axis for both (a) observer and (b) therapist ratings ofcouples.

Observer ratings (PRP)The PRP consists of a single scale, but two principal axes can be isolated for the purposeof the analysis, the DP and the PS. Observer and Therapist ratings of all couples onthe DP and PS scales at the consultation, mid-point of therapy, and the end of thetherapy are shown in Table 1. Repeated measures ANOVA were carried out on theobserver PRP ratings on both the PS and the DP scale of the PRP. The improvement incouple functioning on the DP scale between the beginning and end of the therapy issignificant, F(2,34) = 8.769; p < .001. Pairwise comparisons (with Bonferroni adjustmentfor multiple comparisons) showed that observer rating of couples is significantly differentfrom the beginning of therapy (the consultation) to the end, ES (Consult to End) d =0.92, but that there is no significant difference on the DP scale between the first sessionand the mid-point of therapy.

In other words, though the trend is consistently upwards, the significant changes areregistered only at the end point of the therapy (after 39 sessions), not at the mid-point.


Table 1. Mean scores of couples on DP and PS scales, as rated by therapist and independent rater

Depressive Paranoid schizoid

N = 18 Mean SD Mean SD

OBSERVERConsultation 2.18 (0.62) 2.16 (0.63)Mid-therapy 2.43 (0.80) 2.36 (0.88)End of therapy 2.87 (0.87) 2.83 (0.86)

THERAPISTConsultation 1.94 (0.52) 2.15 (0.60)Mid-therapy 2.40 (0.57) 2.60 (0.82)End of therapy 3.07 (0.78) 3.36 (0.99)

Note. Observer: DP Scale: Consult to End, F(2,34) = 8.769; p � .001; ES d = 0.92.Observer: PS Scale: Consult to End, F(2,34) = 9.273; p � .001; ES d = 0.89.Therapist: DP Scale: Consult to End, F(2,34) = 32.279; p � .001; ES Consult to Mid d = 0.84; Mid toEnd d = 0.98; Consult to End d = 1.7.Therapist: PS Scale: Consult to End, F(2,34) = 16.266; p � .001; ES Consult to Mid d = 0.62; Mid toEnd d = 0.84; Consult to End d = 1.47.NB ES, Effect Size.Effect Sizes are shown without confidence intervals, because this is a small-scale, sample-specific study,not suitable for generalization to a wider population.

The improvement in couple functioning on the PS scale between the beginning andend of the therapy was also significant, F(2,34) = 9.273; p < .001.

Table 1 also shows the overall mean scores of couples on PS scale, as rated by observer.(It is important to point out, that for this scale, items are reverse-scored, meaning thathigher scores on the PS scale mean that the couple were less paranoid schizoid, that is,that they are functioning less in a manner that is associated with psychological ill-health.)

As with the DP scale, pairwise comparisons (with Bonferroni adjustment for multiplecomparisons) on the PS scale showed that observer rating of couples is significantlydifferent from the beginning of therapy (the consultation) to the end, ES (Consult toEnd) d = 0.89, but that there is no significant difference between the first session andthe mid-point of therapy. Therefore, it is important to note that on both the PS and theDP scales the significant differences that were obtained relate to the differences betweenthe beginning and the end of the therapy, that is, to the full 39 sessions.

Therapist ratings of patterns of interpersonal relating (PRP)Repeated measures ANOVA were also carried out on the therapist PRP ratings on boththe PS and the DP Scales of the PRP. As with the observer ratings, significant changeswere found across the 40 session therapies.

On the DP scale, the main effect of time was significant, F(2,34) = 32.279; p <

.001. Unlike the observer ratings, pairwise comparisons (with Bonferroni adjustment formultiple comparisons) showed that therapists were rating the couples as significantlybetter at both the 20-session and 40-session time points. (ES Consult to Mid d = 0.84,Mid to End d = 0.98, Consult to End d = 1.7).

On the PS scale, the main effect of time was also significant, F(2,34) = 16.266;p < .001. Pairwise comparisons (with Bonferroni adjustment for multiple comparisons)showed that therapists were rating the couples as significantly better at both the

Couple therapy 301

20-session and 40-session time points (ES Consult to Mid d = 0.62, Mid to End d =0.84, Consult to End d = 1.47).

Importantly, there were no significant differences between observer and therapistratings when compared directly using repeated measures ANOVA with ‘rater’ as a factor.This was tested for both the PS and DP scales, both across the sessions (time) and alsodirectly compared at each of the three time points.

Hypothesis (2) that participants would show improvement in self-ratings ofpsychological state (CORE) at the end of therapyMean and standard deviation CORE scores for males and females at pre-consultation(before the start of therapy), consultation (beginning), middle, and end of therapy areshown in Table 2. Lower scores indicate improvement in self-ratings of psychologicalstate. Data are corrected to replace missing values (see below).

Because there were equal time periods between the therapy sessions, (i.e., the datawere sampled uniformly throughout), Repeated Measures ANOVA was carried out onthe participants’ CORE scores. This is a 2 × 3 ANOVA model, with factors of gender (2)and time (3), in which both independent variables are within-subjects effects. Time inthis model included the Consult, Mid, and End of Therapy sessions.

The main effect of time was significant, F(2,34) = 6.882; p < .005, with Consultationbeing significantly different from Mid Therapy and End of Therapy. Mid Therapy andEnd of Therapy were not significantly different from one another (ES Consult to MidTherapy: d = 0.53, ES Consult to End Therapy: d = 0.64).

In addition, the main effect of gender was also significant F(1,17) = 5.155; p < .05.However this effect of gender was only found to be significant when running the analysisusing imputed data (see below). Females had significantly lower scores than males atMid and End Therapy (F(1,17) = 5.155; p < .05).

Table 2. Mean and standard deviation CORE and GRIMS scores for males and females at pre-consultation, consultation, mid-point, and end

Male Female

N = 18 Mean SD Mean SD

CORE1

Pre-consultation 1.75 (0.72) 1.19 (0.44)Consultation 1.49 (0.73) 1.30 (0.68)Mid-therapy 1.26 (0.70) 0.96 (0.50)End of therapy 1.27 (0.74) 0.82 (0.62)

GRIMSPre-consultation 42.78 (13.25) 43.17 (11.68)Consultation 41.56 (13.57) 43.83 (9.07)Mid-therapy 39.72 (16.37) 41.44 (15.10)End of therapy 40.50 (15.63) 38.56 (16.57)

Note. 1Time (Consult to End): F(2,34) = 6.882; p � .005.Gender: F(1,17) = 5.155; p � .05.Time × Gender interaction = non-significant.ES Consult to Mid-therapy: d = 0.53.ES Consult to End Therapy: d = 0.64.This table shows data corrected to replace missing values. In both CORE and GRIMS, the tables displayscores of males and females separately.


The interaction of time × gender was non-significant (i.e., both males and femalesshow the same sort of pattern to their scores over the sessions).

The data reported here include a small correction for missing data. Specifically, twocouples had their data repeated for both middle and end points and a further case haddata repeated at both these points for the male partner only. The analyses were carriedout (as above) with the data corrected to replace missing values. Then the analyses wererun again excluding the couples where there was this missing data and, we had the samehighly significant effect of therapy F(2,28) = 6.718; p < .005, but no significant effect ofgender. So, it was only when imputing the missing data in this way that the main effectof Gender was significant.

Baseline periodThe ‘untreated’ period of 3 weeks between the pre-therapy and the consultation allowedus to check for the consistency of the results across this time period, to see whetherthere was natural variation occurring in participants’ scoring on this measure beforetherapy commenced. Repeated Measures ANOVA was carried out on the participants’CORE scores. This is a 2 × 2 ANOVA model, with factors of gender (2) and time (2),in which both independent variables are within-subjects effects. This showed a non-significant main effect of time (F(1,17) = 0.493; p = 0.49). In other words, there is nosignificant spontaneous variance in the CORE scores across the 3-week control periodbefore therapy. There is also no significant main effect of gender (F(1,17) = 4.035;p = .06). There is however a significant interaction of time by gender, (F(1,17) = 8.229;p < .05), with females having significantly higher CORE ratings at pre-consultation thanthe males, but this difference was no longer as high – nor significantly different – by theConsultation time period.

Clinically significant changeThe CORE measure also offers norms that indicate whether individual scores fall withinthe range of a ‘clinical population’ or not, and therefore allow us to look at whetherchanges in participants’ scores indicate ‘clinically significant’ change.

Table 3 shows numbers of men and women falling into the categories of ‘clinical’ or‘non-clinical’ at each of the three assessment points of the study. (Note: Table 3 showsfigures before missing data have been imputed.)

Chi-Square analysis indicated that while the results are in the right direction, they arenon-significant.

Table 3. Numbers of men and women falling into the categories of ‘clinical’ or ‘non-clinical’ at each ofthe three assessment points of the study

Male Female

N = 18 Clinical % Non-clinical % Clinical % Non-clinical %

Pre-consultation 14 (77.8%) 4 (22.2%) 10 (55.6%) 8 (44.4%)Consultation 10 (55.6%) 8 (44.4%) 8 (44.4%) 10 (56.6%)Mid-therapy 6 (37.5%) 10 (62.5%) 4 (23.5%) 13 (76.5%)End of therapy 8 (47.1%) 9 (52.9%) 3 (17.6%) 14 (82.4%)

Females � 2 = 7.865; df = 3; p = .066; Males � 2 = 6.190; df = 3; p = .103.Table 3 shows figures before missing data have been imputed.

Couple therapy 303

Hypothesis (3) that participants would show improvements in their self-ratings ofrelationship satisfaction (the GRIMS)Table 2 shows GRIMS Scores at Pre-consultation, Consultation, Mid-point, and End oftherapy for males and females. As with the CORE analysis, a Repeated Measures ANOVAwas carried out on the participants’ GRIMS scores.

Although the pattern of results on the GRIMS scale indicate a downward trend (inthe direction of improvement), there were no significant effects of ‘time’ or ‘gender’ oncouples’ self-reported marital satisfaction.

Baseline periodAs with the CORE analysis, the ‘untreated’ period of 3 weeks between the pre-therapyand the consultation were tested to check for the consistency of the results across thistime period. Again a 2 × 2 ANOVA model was used, with factors of gender (2) and time(2), in which both independent variables are within-subjects effects. For the GRIMS data,all effects were non-significant. Results of the ANOVA are as follows: gender (F(1,17) =0.05; p = .82); time (F(1,17) = 0.22; p = .64). interaction of time by gender (F(1,17) =1.14; p = .30).

DiscussionOur analyses indicated that results from two out of three of our outcome measures,(PRP and CORE) were significantly improved at the end of the therapy, whilst theresults from the third measure (GRIMS) did not show significant improvement. In otherwords, as we predicted, couples showed improvements in patterns of interpersonalrelating as assessed by independent observer and by therapists at the end of treatment(PRP). Participants also showed self-reported improvements in psychological state asmeasured on CORE at both the 20-session and the end points. But they did not showthe predicted significant differences in pre–post treatment scores in their self-ratedrelationship satisfaction (GRIMS). By contrast, there was no significant change on eitherof the self-report measures (CORE and GRIMS) over the ‘baseline period’, a 3-weekperiod leading up to the consultation point.

ESs of change during treatment on the measure of interpersonal relating of the couples(PRP) from consultation to end of treatment for observers were d = 0.89 and 0.92 forthe DP and PS scales, respectively. For therapist ratings, these figures were d = 1.17 and1.47 on the same two scales. The ES of the changes from beginning to end of the therapyas measured on CORE, of self-rated psychological state was d = 0.64. In the coupletherapy outcome literature, a recent study of the outcome of integrative behaviouralcouple therapy provides a point of comparison. Christensen et al. (2006) report an ES ofd = 0.85 in terms of decreases in couples’ global distress, pre- to post-treatment, whichis smaller than ours on the PRP measure. A large meta-analytic study of couple therapyoutcome (Shadish et al., 1993) found an overall ES of d = 0.60 for 27 studies. Our ESsare equal to, or larger than the ESs reported by them. It should be noted, however, thatour lower ES was for the self-ratings on the CORE measure (and the studies reportedabove were using self-report measures). The above studies were also RCTs rather thannaturalistic studies, as ours was, and the former are usually associated with smaller ESs.

On the PRP measure, observers and therapists were both rating the couples in a waywhich indicates that they were less ‘paranoid schizoid’ and more ‘depressive’ at theend of therapy, to a degree that exceeded what would be expected by chance alone.


According to the conceptual basis of the measure, the changes indicate that couplesshowed improvement in terms of a greater capacity to be insightful, and reflective, andwere less caught up in repeating and re-enacting difficulties in the relationship betweenthem. They made less use of pathological psychological defence mechanisms, and overallshowed more capacity to tolerate differences and separateness, to see each other in away that was less distorted by their own mental state. In other words, at the end oftherapy, couples were rated on both scales of the PRP, in psychodynamic terms, asfunctioning in a manner associated with greater psychological health.

The data for the PRP converge with that of the CORE, which also indicateimprovements in the psychological health and functioning of the individuals in thecouples concerned. It is interesting that individual psychological health improved, butthat this was not reflected in a significant level of improvement in conscious satisfactionwith relationships, as measured on GRIMS. As mentioned earlier, greater psychologicalhealth may involve tolerating more uncomfortable or painful feelings. Being freer to behonest, to express ambivalence, or to disagree creatively, may be felt by some couplesto be harder work and more complex rather than simply to be an improvement. It maybe too that the GRIMS is less sensitive than the CORE measure and would need a biggersample size to pick up significant effects.

In relation to the question of whether improvements were demonstrated after 20sessions and 40 sessions, it is notable that although CORE indicated improvement after20 sessions, the differences only became apparent on the observer PRP ratings after 40sessions of therapy. This is in keeping with results from our earlier unpublished work,where outcome was assessed at 20 sessions, and found to be approaching significance,but not showing the same strong levels of significance found in the current study after 40sessions. It is reasonable to suggest that this change over time is a result of treatment andit implies that length of treatment is important, with 40 sessions providing a sufficienttime span for significant improvements to be detected using this observer-rated measure.We are aware that our use of the penultimate session for our final measures, becauseof the tendency for presenting symptoms to recur at the end of therapy, may place ourresults at some advantage to other studies that have not considered this factor.

Interestingly our conclusion about length of treatment coincides with the view ofone of the pioneers of ‘brief therapy’, David Malan, who for many years used a 40-sessionformat for the individual therapy he studied (Malan, 1963). There is a good theoreticalbase for the view that change, if it is to be lasting, needs to be established and re-established over time, through painful re-working of the understandings reached aboutthe present and the influence upon it of the past. In relation to couples, the two partnersshare the work, sharing the painfulness and the unconscious forces resisting change,as well as the wish to help each other and themselves to develop. This study supportsthe idea that such development can be achieved through time-limited psychoanalyticallybased therapy for couples.

Limitations of the study

CircularityIn relation to the PRP outcomes, it might be suggested that there is a problem ofcircularity: that the qualities being rated by therapists and observers on the PRP measureare what they would expect to see – a reflection of therapists’ expectations andinvestment in the therapy, and of the observers’ prior training and therefore, expectation.This is a difficult charge to refute in a study which is naturalistic, undertaken in a clinical

Couple therapy 305

setting, and which uses an instrument (the PRP) that relies on experienced and well-trained raters. However, the convergent evidence of observer and therapist ratings onthe PRP and participant self-ratings on the CORE gives some confidence for the claim thatwe are tapping evidence of change in the qualities of transactions between the partner ina couple (as assessed on the PRP) and reflected in the self-perceived psychological healthand functioning of the participants (CORE), though not necessarily in their conscioussatisfaction with their relationship (as measured on GRIMS). Care was taken to ensurethat the independent raters were blind to the sequence of any one couple’s ratings; theywere rated in the ‘wrong’ order, and never contiguously, so that differences in ratingsbetween beginning and end of each therapy were not influenced by expectation. Inaddition, the items rated in the PRP are not individually required to be aligned with anoverall formulation to produce a useable score, and the raters are not asked to makesuch a formulation. Thus, vulnerability to a charge of circularity was minimized.

Small sample size and lack of control groupThese limit the inferences that can be drawn from our results and the generalizability ofthe findings. It is possible that, in the absence of a control group, spontaneous remissionmight explain our results, but the finding that untreated control groups of distressedcouples tend to show no improvement, and even get worse (Baucom, Hahlweg, andKuschel, 2003), gives some grounds for seeing our results as indicative evidence of theeffectiveness of psychodynamic couple therapy. These results suggest that it would beworth replicating the study on a larger scale. There are always problems in establishinga control group in naturalistic clinical research studies, because treatment cannot bedenied, but a comparison group could perhaps be offered a shorter and more limitedform of counselling and then be reviewed over the same time period, and comparisonwith other couple therapy treatments could also be made. The strengths of this studyare the combination of self-report measures and observer and therapist ratings, and thefirst ever application of the psychodynamic instrument, the PRP, as a measure of changein couple psychotherapy. Our data provide some encouragement for the utility of thePRP as a measure of change in psychodynamic couple psychotherapy. We suggest thatthis small, naturalistic pilot study may prove a useful model for larger scale studies ofoutcome in psychodynamic couple therapy, which are so needed in this field.

Case examplesIt may be helpful to illustrate these results with a more qualitative glimpse of our workwith the couples. These vignettes have been disguised by altering identifiable details.

One couple in their 40s, with two sons, came into therapy having just finished somebrief counselling elsewhere, which they were unhappy with, saying it was ‘superficial’.The wife was on antidepressants and had also had some sessions of CBT. The therapist’simpression was that the CBT had been used by her to push her into deciding that shemust make concrete changes in her life, one of which was to end the marriage andseparate. At the start of the therapy, the couple were in crisis. They said that for a longtime they had kept up a ‘wall’ between them, which he, in particular, said was a relief.There was a lot of ambivalence about getting into discussing their difficulties with oneanother, but despite an angry stand-off at the start, things did start to open up quitequickly. They became more able to listen to one another, and to allow contact withtheir vulnerability. The work often involved trying to help them to see how quickly they


retreated from this. They became more insightful and more tolerant of their need forhelp over the course of the therapy, and a positive aspect of this was reflected in theirwish to continue, which had to involve referral to colleagues outside the study.

Another couple in their late 40s, with four children, had been living separately forthe best part of a year, though continued to work together. They were stuck in animpasse, unable to separate and coming back together regularly for family meals, work,and meeting up alone together. They had separated following a short-lived affair of his,the year before. Both seemed very thin-skinned, easily wounded or rebuffed, and hidingtheir feelings, giving the impression of being uninterested or unconcerned about theother. When challenged, more depth of feeling was revealed.

She tended to keep herself at a distance, treating her husband rather as though shewere the mother of an exasperating, naughty boy. He had a very disturbed early lifein a Latin American country, separated from his mother and siblings, brought up inwhat was seen as a privileged position, materially, but which was in fact emotionallydeprived. He lived with extended family, not his parents, and he was beaten regularly.At the time of the therapy, he was drinking heavily, and living quite an isolated life, apartfrom his contact with her and with the family. There was some indication of suicidalrisk.

Gradually, the woman became more aware of her usual pattern of holding herselfback, keeping herself emotionally ‘out of sight’, so to speak, whilst complaining of herhusband that he was dominating and controlling. This linked in her mind to her mother,whom she experienced as overbearing and frightening. She showed some insight intohow this position was in itself very controlling, aiming to hold him and others at adistance and controlling what they could know of her. There were several sessionswhere more spontaneous and involved contact seemed to be possible between them.His complaint that she had held herself back throughout their marriage, leaving himfeeling that he had little real contact with her, was expressed more directly. At the sametime, she emerged more from her ‘retreat’ which she associated with the way in which,as a little girl, she had been able to hide herself away in tiny spaces from the rest of thefamily, both to escape volatile rows, but also giving her the feeling of being special andsuperior. As she emerged, he became more anxious about being dominated himself, byher, and that now ‘the boot would be on the other foot’.

As the ending approached, the husband voiced a sense of hopelessness anddepression, complaining that there was no place for him in the family and that hewas treated as unwanted clutter or rubbish. He conveyed that he could not access hisold defences in quite the same way anymore (e.g., a manic state of drinking/socializing).The therapist felt that he was more in touch with his need and vulnerability and thisfelt very difficult for him, when the therapy was ending. His anger about it seemed tolead him into making the therapy into ‘nothing’, of no help. He spoke of anger towardshis wife and their children, who reject him; having invited him to make an effort, to getexcited about an arrangement, they then let him down. He could respond when thiswas explored in terms of his experience of the therapy and the fixed ending, initially anexciting invitation that had turned into a let-down.

Like his wife, he was able to make some sense of his feelings of isolation and hisreadiness to feel pushed out, and linked this to his early experiences. He made thesuggestion that he might have individual therapy, and the therapist left it to him to makecontact after the therapy ended if he wished to pursue this. He did so, and was referred.

Although there was a question about whether the value of the work would be heldonto, there was some comment in the final sessions about how things had improved

Couple therapy 307

between them. The wife pointed out how much they had changed, and gave the exampleof a playful episode at the weekend, in a situation where once, as they both thenrecognized, he would have become very angry. Although this alternated with a muchangrier and more disappointed view, in which they complained they were ‘further apartthan ever’, overall a more mixed and more hopeful picture emerged.

From these case examples, it can be seen that outcomes often do not equate withcrude or simple notions of therapeutic change. It is not a case of ‘happily ever after’.In the current climate of competition between therapeutic modalities, the clinicianmay easily feel that there is little place for recognition of the tremendous difficulty ofeffecting change, its partiality and fragility, and how it may be expressed in such mixedand nuanced ways.

ConclusionsThis relatively small but pioneering study found that couples’ functioning, as measuredby two of our three instruments, improved significantly over time, concurrent with theirhaving 40 sessions of psychodynamic couple therapy. The findings support the use ofthe Couple version of the PRP as an outcome measure, and indicate the likelihood thata larger study using the same instruments and with a control group would confirm theeffectiveness of time-limited psychodynamic couple psychotherapy over 40 sessions.

AcknowledgementThe study was funded by the Department for Education and Schools. We wish to acknowledgethe contributions of the third member of our research team, Francis Grier, who was notavailable to co-author this paper, and also Kate Grayson of Statistics by Design, for the dataanalysis.

ReferencesBarkham, M., Culverwell, A., Spindler, K., Twigg, E., & Connell, J. (2005a). The CORE–OM in an

older adult population: Psychometric status, acceptability and feasibility. Ageing and MentalHealth, 9, 235–245.

Barkham, M., Evans, C., Margison, F., McGrath, G., Mellor-Clark, J., Milne, D., & Connell, J.(1998). The rationale for developing and implementing core batteries in service settings andpsychotherapy outcome research. Journal of Mental Health, 7, 35–47.

Barkham, M., Gilbert, N., Connell, J., Marshall, C., & Twigg, E. (2005b). The suitability and utilityof the CORE-OM and CORE-A for assessing severity of presenting problems in primary andsecondary care based psychological therapy services. British Journal of Psychiatry, 186 ,239–246.

Barkham, M., Mellor-Clark, J., Connell, J., & Cahill, J. (2006). A core approach to practice-basedevidence: A brief history of the origins and applications of the CORE OM and CORE System.Counselling and Psychotherapy Research, 6 , 3–15.

Baucom, D. H., Hahlweg, K., & Kuschel, A. (2003), Are waiting-list control groups needed in futuremarital therapy outcome research. Behaviour Therapy, 34, 179–188.

Baucom, D. H., Shohan V., Mueser, K. T., & Stickle, T. R. (1998). Empirically supported coupleand family interventions for marital distress and adult mental health problems. Journal ofConsulting and Clinical Psychology, 66 , 53–58.

Boddington, S., & Lavender, B. (1995). Treatment models for couples therapy: A review of theoutcome literature and the Dodo’s verdict. Journal of Sexual and Relationship Therapy, 10,69–81.


Bodenmann, G., Plancherel, B., Beach, S. R. H., Widmer, K., Gabriel, B., Meuwly, . . . , Schramm, E.(2008). Effects of coping-oriented couples therapy on depression: A randomized clinical trial.Journal of Consulting and Clinical Psychology, 76(6), 944–954.

Christensen, A., Atkins, D. A., Yi, J., Baucom, D. H., & George, W. H. (2006), Couple and individualadjustment for 2 years following a randomized clinical trial comparing traditional versusintegrative behavioural couple therapy. Journal of Consulting and Clinical Psychology, 74(6),1180–1191.

Christian, J. L., O’Leary, K. D., & Vivian, D. (1994). Depressive symptomatology in maritallydiscordant women and men: The role of individual and relationship variables. Journal ofFamily Psychiatry, 8, 32–42.

Connell, J., Barkham, M., Stiles, W. B., Twigg, E., Singleton, N., Evans, O., & Miles, J. N. V. (2007).Distribution of CORE-OM scores in a general population, clinical cut-off points, and comparisonwith CIS-R. British Journal of Psychiatry, 190, 69–74.

Crowe, M. (1978). Conjoint marital therapy: A controlled outcome study. Psychological Medicine,8, 623–636.

Cummings, E. M., & Davies, P. T. (1994). Children and marital conflict: The impact of familydispute resolution. New York: Guilford.

Dunn R. L., & Schwebel, A. I. (1995) Meta-analytic review of marital therapy outcome research.Journal of Family Psychology, 9, 58–68.

Emmelkamp, P. M. G., Van Der Helm, M., Macgillavry, D., & Van Zanten, B. (1984). Maritaltherapy with clinically distressed couples: A comparative evaluation of system ± theoretic,contingency contracting and communication skill approaches. In K. Hahlweg & N. Jacobson(Eds.), Marital interaction: Analysis and modification. New York: Guilford Press.

Evans, C., Connell, J., Margison, F., & Mellor-Clark, J. (2002), Towards a standardised brief outcomemeasure: Psychometric properties and utility of the CORE-OM. British Journal of Psychiatry,180, 51–60.

Evans, C., Mellor-Clark, J., Margison, F., Barkham, M., McGrath, G., Connell, J., & Audin, K. (2000),Clinical outcomes in routine evaluation: The CORE-OM. Journal of Mental Health, 9(3), 247–255.

Fals-Stewart, W., Klostermann, K., Yates, B. T., O’Farrell, T. J., & Birchler, G. R. (2005), Briefrelationship therapy for alcoholism: A randomized clinical trial examining clinical efficacy andcost-effectiveness. Psychology of Addictive Behaviours, 19(4), 363–371.

Fincham, F. D., Grych, J. H., & Osborne L. N. (1994). Does marital conflict cause child maladjust-ment? Directions and challenges for longitudinal research. Journal of Family Psychology, 8,128–140.

Gilbody, S. M., House, A., & Sheldon, T. A. (2002), Outcomes research in mental health, a systematicreview. British Journal of Psychiatry, 181, 8–16.

Harold, G. T., & Conger, R. D. (1997). Marital conflict and adolescent distress: The role of adolescentawareness. Child Development, 68, 330–350.

Hobson, R. P., Patrick, M. P. H., & Valentine, J. D. (1998). Objectivity in psychoanalytic judgements.British Journal of Psychiatry, 173, 172–177.

Hotopf, M., Churchill, R., & Lewis, G. (1999). Pragmatic randomised trials in psychiatry. BritishJournal of Psychiatry, 175, 217–223.

Kelley, M. L., & Fals-Stewart, W. (2002). Couple- versus individual-based therapy for alcohol anddrug abuse: Effects on children’s psychosocial functioning. Journal of Consulting and ClinicalPsychology, 70, 417–427.

Lanman, M., & Grier, F. (2001). A psychoanalytic approach to brief marital psychotherapy. In F.Grier (Ed.), Brief encounters with couples, some analytical perspectives. London: Karnak.

Lanman, M., & Grier, F. (2001). Evaluating change in couple functioning: A psychoanalyticperspective. Sexual and Relationship Therapy, 18(1), 13–24.

Lanman, M., Grier, F., & Evans, C. (2003). Objectivity in psychoanalytic assessment of couplerelationships. British Journal of Psychiatry, 182, 255–260.

Couple therapy 309

Leff, J., Vearnals, S., Brewin, C. R., Wolff, G., Alexander, B., Asen, E., . . . , Everitt, B. (2000). TheLondon depression intervention trial: Randomised controlled trial of antidepressants versuscouple therapy in the treatment and maintenance of people with depression living with apartner: Clinical outcome and costs. British Journal of Psychiatry, 177, 95–100.

Malan, D. (1963). A study of brief psychotherapy. New York: Plenum.Mellor-Clark, J., Connell, J., Barkham, M., & Cummins, P. (2001). Counselling outcomes in primary

health care: A CORE System data profile. European Journal of Psychotherapy, Counsellingand Health, 4, 65–86.

Moving Forward Together, Lord Chancellors Department (now Department for Children, Schoolsand Families) doc (2002). TSO, Parliamentary Press, Mandela Way, London, SE1 5SS.

National Institute for Health and Clinical Excellence: Depression in Children and Young People(2005). published by the National Collaborating Centre for Mental Health. Standon House,London E1 8AA.

National Institute for Health and Clinical Excellence: The treatment of depression in adults.NICE clinical guideline 90. National Collaborating Centre for Mental Health, DoH | 2009"type="Other">National Institute for Health and Clinical Excellence: The treatment of depres-sion in adults. NICE clinical guideline 90. National Collaborating Centre for Mental Health,DoH (2009), MidCity Place, 71 High Holborn, London, WC1V 6NA.

O’Farrell, T. J., Murphy, C. M., Stephan, S. H., Fals-Stewart, W., & Murphy, M. (2004). Partnerviolence before and after couples-based alcoholism treatment for male alcoholic patients: Therole of treatment involvement and abstinence. Journal of Consulting and Clinical Psychology,72(2), 202–217.

Pape Cowan, C., & Cowan, P. (1997). Working with couples during stressful transitions. In S.Dreman. The family on the threshold of the 21st century: Trends and implications (Ed.),New Jersey, USA and London, UK: Laurence Erlbaum Associates.

Pinsof, W., Wynne, L., & Hambright, A. (1996). The outcomes of couple and family therapy:Findings, conclusions and recommendations. Psychotherapy: Theory, Research, Practice,Training, 33(2), 321–331.

Richardson, P. (2003). Evidence-based practice and the psychodynamic psychotherapies. In S.Mc Pherson, P. Richardson, & P. Leroux (Eds.), Clinical effectiveness in psychotherapy andmental health. London: Karnac.

Rust, J., Bennun, I., Crowe, M., & Golumbok, S. (1990). The GRIMS. A psychometric instrumentfor the assessment of marital discord. Journal of Family Therapy, 12(1), 45–57.

Ruszczynski, S. (1993). Psychotherapy with couples. London: Karnac.Segal, H. (1973). Introduction to the work of Melanie Klein. The International Psycho-analytical

Library, no. 91. M. MK.R. Khan (Ed.). London: Hogarth.Shadish, W. R., Montgomery, L. M., Wilson, P., Wilson, M. R., Bright, I., & Okwumabua, T. (1993).

Effects of family and marital psychotherapies: A meta-analysis. Journal of Consulting andClinical Psychology, 61, 992–1002.

Simpson, L. E., Atkins, D. C., Gattis, K. S., & Christensen, A. (2008). Low-level relationshipaggression and couple therapy outcomes. Journal of Family Psychology, 22(1), 102–111.

Snyder, D. K., & Wills, R. M. (1989). Behavioural versus insight-oriented marital therapy: Effectson individual and interspousal functioning. Journal of Consulting and Clinical Psychology,57(1), 39–46.

Snyder, D. K., Wills, R. M., & Grady-Fletcher, A. (1991). Long-term effectiveness of behaviouralversus insight-oriented marital therapy: A four-year follow-up study. Journal of Counsellingand Clinical Psychology, 59, 138–141.

Received 4 September 2009; revised version received 29 May 2011

An evaluation of time-limited psychodynamic psychotherapy for couples: A pilot study

Documents

Transcript of An evaluation of time-limited psychodynamic psychotherapy for couples: A pilot study