Treating Post-Stroke Emotional Difficulties with a CBT-Informed Approach: A Single Case Experimental Design

What follows is an anonymised case study I submitted during doctoral training in Clinical Psychology. The person this case study is based upon consented to dissemination of the following report. However, all personal identifiable information has been removed or edited.

This paper describes an individual CBT-informed intervention undertaken with someone experiencing emotional difficulties following a stroke. A single case experimental design (SCED) was used to monitor change over time, and to test the effectiveness of the intervention.

Nature of Post-Stroke Emotional Difficulties

Emotional difficulties are common following a stroke (e.g., Barker-Collo, 2007; Whyte & Mulsant, 2002), although exact estimates of post-stroke depression and anxiety vary depending on the study (Whyte & Mulsant, 2002). In addition, some people develop symptoms of post-traumatic stress disorder (PTSD: Favrole et al., 2013; Merriman, Norman, & Barton, 2007; Sembi, Tarrier, O’Neill, Burns, & Faragher, 1998).

Early recognition and treatment of post-stroke emotional difficulties is critical, and a lack of effective treatment can have additional consequences for clients’ physical wellbeing (Lincoln, Kneebone, Macniven, & Morris, 2011). Post-stroke depression can have a negative effect on rehabilitation (e.g., Gillen, Tennen, McKee, Gernert-Dott, & Affleck, 2001) and is associated with increased levels of disability (e.g., Paolucci et al., 2001) and mortality (e.g., House, Knapp, Bamford, & Vail, 2001). Therefore, addressing someone’s emotional difficulties is also important for their physical recovery.

Policy Indicators

The National Institute for Health and Care Excellence (NICE) guidelines are to provide support and education about emotional adjustment after stroke to stroke survivors and their families (NICE, 2014a). Where cognitive difficulties are not present, the suggestion is to manage depression (including adults with a chronic physical health problem: NICE, 2014b) and anxiety (NICE, 2011) in line with the recommendations for adults. The recommendations for adults with anxiety or persistent depression include providing group treatment, computerised cognitive behaviour therapy (CBT), or individual guided self-help based on CBT (NICE, 2014c). In addition, the guidelines for post-traumatic stress disorder suggest trauma-focused CBT, with longer sessions when discussing the traumatic event itself (NICE, 2012).

Single Case Experimental Design

The SCED helps to bridge the gap between rigorousness and relevance, which are often thought to be at odds(Barkham & Mellor-Clark, 2000; McMillan & Morley, 2010; Parry, 2000). By focusing on an individual client, the SCED closely matches routine clinical practice, and allows clinicians to answer questions about process that cannot easily be answered by group-level data. It also allows clinicians to assess the efficacy of a treatment for a client who differs from the clients involved in forming the evidence base, thus adding to its clinical relevance. At the same time, in order to maintain the validity of the study, three key strategies are used. 1) Repeated measurement allows the researcher to assess the extent to which the variation is consistent with their hypothesis or with a plausible alternative. 2) A baseline period containing repeated measurements allows assessment of the clinical issue prior to intervention. A stable baseline suggests that any significant change during the intervention phase is related to the intervention in some way. 3) Experimental manipulation adds to the validity of a SCED by randomly allocating periods of time to a particular treatment or lack of treatment, thus helping to rule out alternative explanations for changes.

There is currently only a limited evidence base to suggest that CBT is helpful for people experiencing emotional difficulties following a stroke (Lincoln et al., 2011), and the majority of this has focused on post-stroke depression. Therefore, it would be useful to evaluate the effectiveness of CBT with a client who has several co-morbid difficulties.

Case Overview

The client presented within a community stroke service with a combination of anxiety, depression, and intrusive memories following her stroke. These symptoms were impacting on her day-to-day functioning and her physical rehabilitation.

Hypotheses

The present study aimed to evaluate the effectiveness of a CBT-informed intervention in reducing frustration, symptoms of anxiety, and the impact of intrusive memories, whilst increasing emotional wellbeing. Hypotheses included:

There will be a reduction in the client’s symptoms of depression and anxiety, as measured by a combination of idiographic and standardised measures.
The client will be less bothered by her intrusive memories, as measured by idiographic ratings.
The client will report higher levels of happiness and overall functioning, as measured by a combination of idiographic and standardised measures.

Method

Design

An A/B SCED was chosen for this study. Repeated measurements were taken during a two-week baseline assessment period (A) and during an eight-week CBT-informed intervention (B). In total, measurements were taken over the course of 62 consecutive days, although there were four days where the participant did not complete measures during the intervention period, leaving 58 days of repeated measurements (14 days baseline).Symptoms were measured through a combination of standardised measures and daily client recordings. The effectiveness of the intervention was assessed by measuring change in the client’s self-reported symptoms of anxiety, depression, and trauma, relative to these symptoms during the assessment period.

Participant

Grace (pseudonym), is a 54 year old white British female, who was referred for psychology input by her Occupational Therapist (OT) within the community stroke service. Grace’s OT reported that Grace seemed anxious following her stroke.

In the initial session, Grace reported feelings of low mood, anxiety, lack of motivation, low energy levels, and poor sleep. She also reported having intrusive memories and trying to avoid anything that reminded her of her stroke. She described feeling that she was unable to do the things that she used to do, and feeling very frustrated since her stroke. She also described having a “knotted feeling” in her stomach and experiencing intrusive thoughts and images related to her stroke, which interfered with her daily activities and with her sleep.

Measures

Idiographic measures. Four individualised measures were designed collaboratively with the client to capture the specific difficulties the client was experiencing. These measures are presented in Table 1 (see Appendix A for an example self-monitoring form).

Table 1.Idiographic measures designed collaboratively with client

	Client statement	Frequency	Scale
ID1	“Feeling frustrated”	Daily	0 ‘not at all– 10 ‘a lot’
ID2	“Knotted feeling intensity”	Daily	0 ‘not at all– 10 ‘a lot’
ID3	“Being bothered by memories of the stroke”	Daily	0 ‘not at all– 10 ‘a lot’
ID4	“Feeling happy”	Daily	0 ‘not at all’– 10 ‘a lot’

All of the idiographic measures were designed to have a 0-10 Likert scale. For all of these items except “feeling happy,” a higher score indicated more severe symptoms.

Nomothetic measures. Three standardised measures were included. The Patient Health Questionnaire (PHQ-9: Kroenke & Spitzer, 2002) and the Generalized Anxiety Disorder (GAD-7: Spitzer, Kroenke, Williams, & Löwe, 2006) were used at three key time points: the beginning of the first assessment session, the end of the final assessment session, and the end of the intervention. In addition, the Outcome Rating Scale (ORS: Miller & Duncan, 2003) was used at the end of every session.

The PHQ-9 is a self-report questionnaire that asks clients to rate the frequency of their symptoms over the past two week period. There are nine items rated on a four point scale (not at all, several days, more than half the days, or nearly every day), with a maximum score of 27. A large validation study (Kurt Kroenke, Spitzer, & Williams, 2001) included 6,000 participants recruited from primary care and obstetrics-gynaecology outpatient clinics (n=3,000 for each). Normative data were reported for people diagnosed with major depression (n=41, M=17.1, SD=6.1), other depressive disorder (n=65, M=10.4, SD=5.4), and no depressive disorder (n=474, M=3.3, SD=3.8).Internal consistency (Cronbach’s α=0.89) and test-retest reliability (r=.84) were high.A pre-treatment score of 10 or above on the PHQ-9 has been suggested as a cut-off score to indicate depression (The IAPT Data Handbook, 2011). The PHQ-9 has also been validated for use with people who have had a stroke (Williams et al., 2005a), and a cut-off score of 10 or above was found to have good sensitivity and specificity for this client group as well (91% sensitivity and 89% specificity for major depression; 78% sensitivity and 96% specificity for any depression diagnosis).

The GAD-7 follows the same format as the PHQ-9, but it includes seven rather than nine questions, and focuses on anxiety rather than depression. The maximum score is 21, and a pre-treatment score of 8 or above is the suggested clinical cut-off for generalised anxiety (The IAPT Data Handbook, 2011). No stroke-specific cut-off score has been suggested. A large validation study (Spitzer et al., 2006)included 2,740 participants recruited from primary care clinics in the United States. Normative data was given for people diagnosed as having generalised anxiety disorder (n=73, M=14.4, SD=4.7) and people without generalised anxiety disorder (n=892, M=4.9, SD=4.8). Internal consistency (Cronbach’s α=0.92) and test retest reliability (r=.83) were both high.

The ORS(Miller & Duncan, 2004)is a four-item visual analogue scale that asks people to rate how they have been feeling during the previous one week period. The four items include: 1) overall, 2) individually, 3) interpersonally, and 4) socially. This additional measure provides a way of monitoring change on a session-by-session basis and is concise to complete. A validation of the ORS (Miller & Duncan, 2003) included 521 participants recruited from a community family service agency. Non-clinical participants included masters level students, therapists, and staff (n=87), whilst clinical participants (n=435) had presented with a range of difficulties and were attending outpatient counselling. Normative data was reported for the clinical group (M=19.6, SD=8.7) and the non-clinical group (M=28, SD=6.8). Internal consistency (coefficient alpha across four administrations=0.93) and test retest reliability (r=0.66 for the first two administrations, decreasing with each subsequent administration) were calculated using data from the non-clinical sample. For the ORS, higher scores indicate better functioning, with a maximum score of 40, and a suggested clinical cut-off of 25 (Miller & Duncan, 2004).

Clinical formulation

Given the links between Grace’s physical and emotional difficulties following her stroke, the formulation focused on how the biological vulnerability factors from her stroke (and the impact of those factors) fed into cognitive behaviour therapy (CBT) maintenance cycles. INFORMATION REDACTED FOR ANONYMITY. The CBT maintenance cycles were developed collaboratively in the sessions with Grace, and their relationship with the precipitating vulnerability factors were considered throughout our work. This formulation does not include information about Grace’s past, as this was not the main focus of our work. However, these factors were considered in supervision.

Intervention

A CBT-informed approach (e.g., Westbrook, Kennerley, & Kirk, 2011) was used, and sessions were 90 minutes for talking about Grace’s experience of her stroke (in line with: NICE, 2012) and to allow extra time given Grace’s communication difficulties. The intervention consisted of four primary components which were implemented across eight sessions (9 sessions were planned, but 1 was cancelled):

Psychoeducation (with Grace and her wife)
Behavioural activation
Graded exposure
Challenging thoughts

Ethics

The use of the PHQ-9 and GAD-7 were standard practice within the service. The data is used to provide feedback to the team about individual clients’ psychological wellbeing, and to trigger referrals if necessary. The possibility of the data being used as part of the trainee’s academic work was discussed with Grace from the first session and she provided informed consent. She also consented to the report being disseminated publicly to help others in training, as long as no personal identifiable information was included. No significant cognitive impairments were identified following Grace’s stroke, and there was no indication that Grace’s capacity to consent to self-evaluation was limited. Daily monitoring was not standard practice within the service, but this data were used clinically within the sessions and not purely for the purposes of the SCED. Treatment was not withheld at any point.

Analytic methods

A combination of graphical and statistical analysis methods were used (as suggested by: McMillan & Morley, 2010; Morley & Adams, 1991) to compare scores during the baseline (control) phase with scores in the intervention (experimental) phase. The analyses are shown separately for idiographic and nomothetic data.

Idiographic. Daily scores for Phase A and B are presented visually with trendlines to assess stability during the baseline period and to allow for comparison with scores in the intervention phase. Two methods were used to assess the degree of overlap of data points between baseline and intervention phases. The percentage of non-overlapping data (PND: Scruggs, Mastropieri, & Casto, 1987), which calculates the percentage of data points in the intervention phase better than the best score in the baseline phase. The percentage of data exceeding the median (PEM: Ma, 2006) was also used because it is not as sensitive to outliers as PND. For PEM, the percentage of scores in the intervention phase better than the median of the baseline phase scores is calculated.

Descriptive statistics are also provided for the baseline and intervention phases, including analysis of variability. To smooth the data and to account for missing data points, running medians were calculated for the idiographic measures during the intervention phase (medians of three: Morley & Adams, 1991) and these are also displayed visually.

In order to account for the issue of serial dependency in time series data, autocorrelation was carried out. If data are serially dependent, then the results must be interpreted with caution.

Nomothetic. In order to assess improvement in the nomothetic measures, the data were presented visually to show the baseline and intervention phases, and descriptive statistics were provided. In addition, reliable and clinically significant change was calculated.

Reliable change assesses whether a client has changed more than could be expected based on measurement error or natural variability. This can be calculated using the Reliable Change Index (RCI: Jacobson & Truax, 1991). Published RCI values for the PHQ-9, GAD-7, and ORS were used in this SCED.

Clinical change assesses whether the client is likely to have moved from a ‘clinical’ to a ‘normal’ range (Bauer, Lambert, & Nielsen, 2004; Jacobson, Follette, & Revenstorf, 1984; Jacobson & Truax, 1991). Published clinical cut-offs were available for each measure, and the participant’s scores were compared to cut-off scores to determine whether they had moved from the clinical to non-clinical range. For both types of analysis, only two scores can be compared at a time. Therefore, scores at the beginning of the assessment period were compared to scores at the end of the assessment period, and scores at the end of the assessment period were compared to scores at the end of the intervention period.

Results

Data will be presented first for the idiographic measures, followed by nomothetic measures. On the graphs, scores for the ORS and “feeling happy” have been reversed so that a higher score indicates worse functioning across all measures. In the graphs and analysis, days when the client did not complete ratings were included as empty data points. Looking at the slopes of the trendlines (see Appendix C) showed that including these dates did not have a noticeable impact on the trendlines.

Idiographic

Visual presentation of data. Idiographic data for both the baseline and intervention phases are presented in Figure 2. The baseline and intervention phases are separated by a vertical line, and separate trendlines are shown for each phase.

Visual inspection of the graphs suggests that the scores during the baseline period were relatively stable for the first three idiographic measures, with a very slight trend toward improvement. There was more variability in the client’s daily ratings of “feeling happy” than for the other measures, and there was a slight trend toward deterioration.

During the intervention phase, the graphs suggest a slight decrease in scores (i.e., improvement) for the first three variables, and very slight increase (i.e., deterioration) in the client’s scores for “feeling happy”. However, the changes are very small, and the trends in the intervention phases (for each individual measure) shift in the same direction as the trends in the baseline phases.

Figure 2. Visual presentation of idiographic data over baseline and intervention phases

Figure 2.3 Graph showing baseline and intervention data for ID3

Figure 2.4 Graph showing baseline and intervention data for ID4

Non-overlap. PND and PEM were calculated for each of the four idiographic measures. The values are set out below in Table 3 (calculations in Appendix E).

Table 3. Percentage of non-overlap of data for idiographic measures
Measure	PND (%)	PEM (%)
Feeling frustrated	22.73	47.73
Knotted feeling intensity	18.18	70.45*
Being bothered by memories of the stroke	13.64	29.55
Feeling happy	0.00	45.45
Note: *=moderately significant effect

For each measure, the results for PND and PEM vary widely, and PND is not significant for any measure. However, given that PEM is less sensitive to outliers and this data set shows a high degree of variability, PEM is likely to be a more accurate measure (Ma, 2006). Looking at PEM on its own, there appears to be a moderate improvement in “knotted feeling intensity” during the intervention phase, with no significant change in the other measures.

Descriptive Statistics. Descriptive statistics for the baseline and intervention phases for each of the four idiographic variables are given in Table 2.In line with the visual analysis, the descriptive statistics suggest very slight improvement from baseline to intervention for all of the variables other than feeling happy, which shows slight deterioration.

Table 2. Descriptive statistics for the baseline and intervention phases
Measure	Baseline Mean	Baseline Standard Deviation	Intervention Mean	Intervention Standard Deviation
Feeling frustrated	9.00	0.78	8.52	1.27
Knotted feeling intensity	9.14	0.95	8.70	1.13
Being bothered by memories of the stroke	8.86	0.86	8.80	1.11
Feeling happy	6.57	2.85	6.80	2.18

Variability. The standard deviations above provide a measure of variability. For the first three idiographic measures, the standard deviations are quite small (<1) in the baseline phase, and they become slightly bigger during the intervention phase. This suggests more variability in the intervention than in the baseline phase. However, for “feeling happy”, the standard deviation in the baseline phase suggests a higher degree of variability, with the standard deviation decreasing slightly in the intervention phase. Plots to illustrate how much each data point in the intervention deviates from the baseline median (Morley & Adams, 1991)are included in Appendix D.

Smoothed intervention data. In order to smooth the data due to missing data and variability, medians of three(Morley & Adams, 1991) were calculated and plotted visually. Figure 3 shows this data, accompanied by trendlines.

Figure 3. Visual presentation of smoothed idiographic data over intervention period

With the data presented in this format, the trendlines show slight improvement for the first two and slight deterioration for latter two of the idiographic measures. For “being bothered by memories of the stroke”, the trend is in the opposite direction from when the raw data are presented without smoothing, which suggests that the analytic methods have an effect on the results (see Discussion).

Autocorrelation. Autocorrelation analysis was carried out to identify serial dependency between the data points (plots shown in Appendix F and Appendix G for baseline and intervention phases, respectively). For the baseline phase, data were only serially dependent for ID3 (all except lag two, p<0.05). However, during the intervention period, autocorrelations were significant for lags in all four of the variables (p<0.05 for ID1: all except lags 10 and 11; ID2: lag 1 only; ID3: lags 1 to 11 and 29; ID4: lag 1 only), indicating serial dependency. Therefore, the results must be interpreted with caution.

Nomothetic

For each of the three nomothetic measures, data are presented visually and descriptive statistics for the baseline and intervention phases are provided. Lower scores on the graphs indicate improvement across all of the measures. Reliable and clinically significant change is calculated and discussed following visual presentation of the data.

Visual presentation. Figure 5 shows the client’s PHQ-9 scores graphically. Scores are shown for the beginning of assessment, end of intervention, and end of intervention.

Figure 5. PHQ-9 scores at three time points

For the PHQ-9, there is no change during the assessment or intervention period. The client’s score remained constant at 17.

Figure 6shows the client’s GAD-7 scores graphically. Scores are shown for the beginning and end of assessment and end of intervention.

Figure 6. GAD-7 scores at three time points

Visual inspection of Figure 6 suggests that the client’s anxiety got better during the assessment period and worse during the intervention. Further analysis is needed to determine whether these changes are significant.

Figure 7 shows the client’s ORS scores graphically. The ORS was completed in each session (approximately weekly).

Visual inspection of Figure 7 suggests that the client might have been improving slightly during the baseline phase. There is an overall trend toward improvement in the intervention phase, but a high degree of variability (standard deviation=7.56 in the intervention phase).

Figure 7. Session-by-session scores on the ORS

Reliable and clinically significant change. RCI and clinically significant change calculations can only take into account two scores at a time. Therefore, for each measure, change was assessed separately between the 1) beginning and end of assessment, 2) end of assessment and end of intervention, and 3) beginning of assessment and end of intervention. Scores for each measure at these three time points are shown in Table 4, along with RCI and clinical cut-off values. For reliable and clinical change calculations, scores on the ORS were not reversed. Therefore, the scores for the ORS given in the tables will not match scores shown on the graphs.

Table 4. Client scores and values for clinical and reliable change
	Score Time 1	Score Time 2	Score Time 3	RCI	Clinical Change Cut-off
PHQ-9	17	17	17	6	10
GAD-7	16	11⁺	13	5	8
ORS¹	17	24⁺	34*⁺	5	25
Note:¹= opposite direction of scale; * = clinically significant change from previous score; ⁺ = reliable change from previous score

The clinical cut-off for the PHQ-9 is a score of 10 (The IAPT Data Handbook, 2011; including for people who have had a stroke: Williams et al., 2005b) and the RCI is 6(The IAPT Data Handbook, 2011). However, the client’s PHQ-9 score stayed constant at 17 at each of the three time points. Therefore, there was no reliable or clinically significant change on the PHQ-9 at any point, and the client remained within the moderately severe range for depression.

The clinical cut-off for the GAD-7 is 8 and the RCI is 5 (The IAPT Data Handbook, 2011). The client’s score started off in the clinical range, and never moved below 8. Change from the beginning to the end of assessment was 5 (i.e., 16-11), which is equal to the RCI, suggesting a borderline reliable change during the baseline phase. Changes from the beginning of baseline to the end of the intervention (i.e., 16-13) and from the end of baseline to the end of the intervention (i.e., 11-13) were less than 5. Therefore, there was no reliable change during the intervention phase. Taken together, the reliable and clinical change calculations suggest that there was some reliable improvement during the baseline phase, but no reliable change during the intervention phase, and none of the changes reached clinical significance.

The clinical cut-off for the ORS is a score of 25 (higher than 25 is considered non-clinical), and the RCI is 5 (Miller & Duncan, 2004). Change from the beginning to the end of assessment was 7 (i.e., 24-17), suggesting a reliable but not clinically significant change during the baseline phase. In addition, the changes from the beginning of assessment to the end of intervention (i.e., 34-17=17) and from the end of the assessment to the end of the intervention (i.e., 34-24=10) were both greater than the RCI of 5, suggesting reliable change overall, and additional reliable change during the intervention period. This also moved the client from the clinical to the non-clinical range during the intervention phase, so the change was both reliable and clinically significant.

Discussion

The results of this SCED suggest that the CBT-informed intervention helped to alleviate some of Grace’s difficulties, but not all of them. In spite of some small improvements, none of the three hypotheses were fully supported by the data. In terms of hypothesis one, the participant’s level of depression did not decrease at all, and her daily ratings of frustration did not decrease significantly. Although her overall anxiety levels decreased reliably during the assessment, they did not decrease during the intervention. However, during the intervention, there was a moderate reduction in one specific symptom of anxiety (“knotted feeling intensity”). For hypothesis two, there was no significant reduction in her daily ratings of being bothered by memories of her stroke. For hypothesis three, the participant did report improvements in her overall functioning, which reached both reliable and clinically significant change levels during the intervention. However, in spite of this, there was no significant change in her ratings of feeling happy.

These results are only partially consistent with current guidance suggesting that CBT can be helpful for people with emotional difficulties following a stroke (e.g., Lincoln et al., 2011; NICE, 2014a). Therefore, using a SCED was useful for looking specifically at which areas benefited from treatment and which did not. In particular, Grace’s level of depression remained entirely unchanged throughout the assessment and intervention periods. There are many possible reasons for this, but the two most plausible reasons are: 1) the intervention was primarily targeting Grace’s stroke-specific anxiety and 2) given Grace’s history of depression, her scores on the PHQ-9 might have pre-dated her stroke, and a short, stroke-focused intervention might not have been sufficient for change. Although her overall level of anxiety did not improve during the intervention, one specific symptom of anxiety did improve a moderate amount. In addition, her overall level of functioning improved. This suggests that the CBT-informed approach used in this case had a limited effect on anxiety and overall functioning, but not depression or symptoms of PTSD.

Limitations

Even the modest results described above need to be interpreted with caution. There is a high degree of variability in the data, which makes analysis (particularly for reliable and clinical change) less meaningful. In addition, the directions of the trends in the baseline phase for each measure are consistent with the directions of the overall trends for each measure. This change during the baseline phase is difficult to avoid, as the very act of beginning treatment can be therapeutic for some people (McMillan & Morley, 2010). For most measures, change during the baseline phase was not significant, but there was reliable change in the baseline phase for both the GAD-7 and the ORS, making it difficult to attribute changes in these measures entirely to the intervention. For the purposes of this SCED, it would have been useful to extend the baseline phase, but this was not practical because of limited time on placement, and because it would have been unethical to withhold treatment, given that the clinical assessment had come to an end. In addition, idiographic data in the intervention phase were found to be serially dependent, so results must be treated with caution.

An additional complication in interpreting the data from this SCED is that the results vary depending on how they are analysed. The results above are for the raw scores, but when medians of three were calculated and plotted, one of the trendlines went in the opposite direction. In addition, the non-overlap analyses were difficult to interpret as the results for PND and PEM were not consistent. This shows the impact that the analysis methods have on the data and its interpretation.

Scores also differ between measures. For example, there is no reliable or clinically significant improvement on the GAD-7 during the intervention phase, and in fact, the trend is toward deterioration. However, during the same period, there is some improvement in the client’s primary symptom of anxiety (which she described as a “knotted feeling”). Although these two things are not exactly the same, one might expect them to be related. This highlights the importance of using measures which are specific to the difficulties being targeted and sensitive to change. On the other hand, it is important to use measures which are robust enough to prevent unnecessary variability. For example, “feeling happy” shifts on a daily basis.In hindsight, although it was good to design the measures collaboratively with the client, it would have been useful to include several other measures. In particular, it would have been helpful to have a standardised measure of trauma/PTSD, and it would have been useful to capture change in the client’s behaviours. These are both areas that the client verbally reported changing, but that were not fully captured by the measures that were used.

There was not time to carry out a follow-up session with the client, which means that it was not possible to assess the long-term effectiveness of the intervention. It would have been useful to find out whether the small gains were lost, maintained, or improved upon. In addition, introducing a period without, or separating the treatment components into different time periods, would have added to the strength of the SCED by introducing further experimental manipulation.

Summary

This SCED aimed to measure the outcome of a CBT-informed intervention for someone experiencing depression, anxiety, and symptoms of PTSD following a stroke. Although the analyses did not fully support the hypotheses set out in the Introduction, the results did indicate that the intervention had a positive effect on a particular symptom of anxiety and on the client’s overall functioning. There was no change in depression, overall anxiety, or being bothered by memories of the stroke. Further research is needed to understand which specific aspects of CBT are helpful for specific types of post-stroke emotional difficulties, including depression, anxiety, and PTSD symptoms.

References

Barker-Collo, S. L. (2007). Depression and anxiety 3 months post stroke: prevalence and correlates. Archives of Clinical Neuropsychology : The Official Journal of the National Academy of Neuropsychologists, 22(4), 519–31. doi:10.1016/j.acn.2007.03.002

Barkham, M., & Mellor-Clark, J. (2000). Rigour and relevance: Practice-based evidence in the psychological therapies. In N. Rowland & S. Goss (Eds.), Evidence-based counselling and psychological therapies: Research and applications (pp. 127–144). London, UK: Routledge.

Bauer, S., Lambert, M. J., & Nielsen, S. L. (2004). Clinical significance methods: a comparison of statistical techniques. Journal of Personality Assessment, 82(1), 60–70. doi:10.1207/s15327752jpa8201_11

Favrole, P., Jehel, L., Levy, P., Descombes, S., Muresan, I.-P., Manifacier, M.-J., & Alamowitch, S. (2013). Frequency and predictors of post-traumatic stress disorder after stroke: a pilot study. Journal of the Neurological Sciences, 327(1-2), 35–40. doi:10.1016/j.jns.2013.02.001

Gillen, R., Tennen, H., McKee, T. E., Gernert-Dott, P., & Affleck, G. (2001). Depressive symptoms and history of depression predict rehabilitation efficiency in stroke patients. Archives of Physical Medicine and Rehabilitation, 82(12), 1645–9. doi:10.1053/apmr.2001.26249

House, A., Knapp, P., Bamford, J., & Vail, A. (2001). Mortality at 12 and 24 Months After Stroke May Be Associated With Depressive Symptoms at 1 Month. Stroke, 32(3), 696–701. doi:10.1161/01.STR.32.3.696

Jacobson, N. S., Follette, W. C., & Revenstorf, D. (1984). Psychotherapy outcome research: Methods for reporting variability and evaluating clinical significance. Behavior Therapy, 15(4), 336–352. doi:10.1016/S0005-7894(84)80002-7

Jacobson, N. S., & Truax, P. (1991). Clinical significance: A statistical approach to defining meaningful change in psychotherapy research. Journal of Consulting and Clinical Psychology, 59(1), 12–19.

Kroenke, K., & Spitzer, R. L. (2002). The PHQ-9: A new depression diagnostic and severity measure. Psychiatric Annals, 32(9), 509–515.

Kroenke, K., Spitzer, R. L., & Williams, J. B. W. (2001). The PHQ-9: validity of a brief depression severity measure. Journal of General Internal Medicine, 16(9), 606–613. doi:10.1046/j.1525-1497.2001.016009606.x

Lincoln, N.B., Kneebone, I.I., Macniven, J.A.B, & Morris, R.C. (2011).Psychological management of stroke. Chichester, United Kingdom: John Wiley & Sons, Ltd.

Ma, H.-H. (2006). An alternative method for quantitative synthesis of single-subject researches: percentage of data points exceeding the median. Behavior Modification, 30(5), 598–617. doi:10.1177/0145445504272974

McMillan, D., & Morley, S. (2010). Single Case Quantitative Methods for Practice-Based Evidence. In M. Barkham, G. E. Hardy, & J. Mellor-Clark (Eds.), Developing and Delivering Practice-Based Evidence: A Guide for the Psychological Therapies (pp. 109–139). Chichester, UK: JohnWiley & Sons Ltd.

Merriman, C., Norman, P., & Barton, J. (2007). Psychological correlates of PTSD symptoms following stroke. Psychology, Health & Medicine, 12(5), 592–602. doi:10.1080/13548500601162747

Miller, S. D., & Duncan, B. L. (2003). The outcome rating scale: a preliminary study of the reliability, validity, and feasibility of a brief visual analogm measure. Journal of Brief Therapy, 2(2), 91–100.

Miller, S. D., & Duncan, B. L. (2004). The outcome and session rating scale: Administration and scoring manual. Chicago, IL.

Morley, S., & Adams, M. (1991). Graphical analysis of single-case time series data. British Journal of Clinical Psychology, 30(2), 97–115. doi:10.1111/j.2044-8260.1991.tb00926.x

National Institute for Health and Clinical Excellence (2011).Post-traumatic stress disorder (PTSD). Retrieved July 15, 2014 from www.nice.org.uk.

National Institute for Health and Clinical Excellence (2012).Generalised anxiety disorder overview.Retrieved July 15, 2014 from www.nice.org.uk.

National Institute for Health and Clinical Excellence (2014a). Assessment and therapy in

specific areas for people with stroke. Retrieved June 10, 2014 from http://pathways.nice.org.uk/pathways/stroke.

National Institute for Health and Clinical Excellence (2014b). Step 1: Recognition, assessment and initial management of depression in adults. Retrieved June 10, 2014 from http://pathways.nice.org.uk/pathways/depression.

National Institute for Health and Clinical Excellence (2014c). Step 3: Persistent subthreshold depressive symptoms or mild to moderate depression with inadequate response to initial interventions, and moderate and severe depression in adults. Retrieved June 10, 2014 from http://pathways.nice.org.uk/pathways/depression.

Paolucci, S., Antonucci, G., Grasso, M. G., Morelli, D., Troisi, E., Coiro, P., … Bragoni, M. (2001). Post-Stroke Depression, Antidepressant Treatment and Rehabilitation Results. Cerebrovascular Diseases, 12(3), 264–271. doi:10.1159/000047714

Parry, G. (2000). Evidence based psychotherapy: special case or special pleading? Evidence-Based Mental Health, 3(2), 35–37. doi:10.1136/ebmh.3.2.35

Scruggs, T. E., Mastropieri, M. A., & Casto, G. (1987). The Quantitative Synthesis of Single-Subject Research: Methodology and Validation. Remedial and Special Education, 8(2), 24–33. doi:10.1177/074193258700800206

Sembi, S., Tarrier, N., O’Neill, P., Burns, A., & Faragher, B. (1998). Does post-traumatic stress disorder occur after stroke: a preliminary study. International Journal of Geriatric Psychiatry, 13(5), 315–22.

Spitzer, R. L., Kroenke, K., Williams, J. B. W., & Löwe, B. (2006). A brief measure for assessing generalized anxiety disorder: the GAD-7. Archives of Internal Medicine, 166(10), 1092–7. doi:10.1001/archinte.166.10.1092

The IAPT Data Handbook. (2011). London, UK.

Westbrook, D., Kennerley, H., & Kirk, J. (2011). An introduction to cognitive behaviour therapy: Skills and applications. (2nd ed.). London, UK: Sage Publications.

Whyte, E. M., & Mulsant, B. H. (2002). Post stroke depression: epidemiology, pathophysiology, and biological treatment. Biological Psychiatry, 52(3), 253–264. doi:10.1016/S0006-3223(02)01424-5

Williams, L. S., Brizendine, E. J., Plue, L., Bakas, T., Tu, W., Hendrie, H., & Kroenke, K. (2005a). Performance of the PHQ-9 as a screening tool for depression after stroke. Stroke; a Journal of Cerebral Circulation, 36(3), 635–8. doi:10.1161/01.STR.0000155688.18207.33

Williams, L. S., Brizendine, E. J., Plue, L., Bakas, T., Tu, W., Hendrie, H., & Kroenke, K. (2005b). Performance of the PHQ-9 as a screening tool for depression after stroke. Stroke: A Journal of Cerebral Circulation, 36(3), 635–638.

Appendix A

Sample client rating sheet

Please rate each of the following on a scale of

0 (not at all) to 10 (a lot).

	Feeling frustrated	Knotted feeling intensity	Being bothered by memories of the stroke	Feeling happy
Friday, 01/08/14
Saturday, 02/08/14
Sunday, 03/08/14
Monday, 04/08/14
Tuesday, 05/08/14
Wednesday, 06/08/14
Thursday, 07/08/14

Missing Data Trendline Slope Analysis for Intervention Phase

Table C. Trendline slope analysis
	Dates included (slope of trendline)	Dates deleted (slope of trendline)	Difference
Feeling frustrated	-0.029	-0.004	0.025
Being bothered by memories of the stroke	0.0018	-0.002	0.004
Knotted feeling intensity	0.0236	0.0276	0.004
Feeling happy	0.0304	0.031	0.001

Appendix D

Plots illustrating deviations from the median for each idiographic measure

Appendix E

Non-overlap calculations

Table D.1 PND Calculations
	Baseline-lowest point	Intervention – number of lower points	Calculation	PND
Feeling frustrated	8	10	(10/44)*100	22.73%
Being bothered by memories of the stroke	8	8	(8/44)*100	18.18%
Knotted feeling intensity	8	6	(6/44)*100	13.64%
Feeling happy	1	0	(0/44)*100	0.00%

Table D.2 PEM Calculations
	Baseline median	Intervention – number of lower points	Calculation	PND
Feeling frustrated	9	21	(21/44)*100	47.73%
Being bothered by memories of the stroke	9.5	31	(31/44)*100	70.45%
Knotted feeling intensity	9	13	(13/44)*100	29.55%
Feeling happy	6.5	20	(20/44)*100	45.45%

Appendix F

Baseline data autocorrelation plots

Intervention data autocorrelation plots