| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
ORIGINAL ARTICLES |
From Williams LifeSkills, Inc. (E.D.K., V.P.W., M.C.H.), Durham, North Carolina; and the Department of Psychiatry and Behavioral Sciences (J.D.L., R.B.W.), Duke University Medical Center, Durham, North Carolina.
Address correspondence and reprint requests to Redford B. Williams, MD, Behavioral Medicine Research Center, 2212 Elder Street, Box 3926, Duke University Medical Center, Durham, NC 27710. E-mail: redfordw{at}duke.edu
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONAPPENDIXNOTESREFERENCES |
|---|
Methods: One hundred ninety-six participants screening positive for elevated psychosocial distress were randomized to either a waitlist control group or one of three intervention groups: the LifeSkills Workshop, the LifeSkills Video, or the LifeSkills Video and Workshop combined. Psychosocial risk factors were evaluated at baseline and at 10 days, 2 months, and 6 months after the training/wait period.
Results: At 10 days follow up, the workshop + video and video-only groups showed significant improvements over control subjects in trait anxiety and perceived stress. Moreover, the workshop + video group maintained benefit over control subjects throughout 6 months follow up in both of these measures, whereas the video-only group maintained benefit in trait anxiety.
Conclusions: Because the psychosocial well-being of two of the treated groups improved over that of the control group, it appears that the Williams LifeSkills program accelerates and maintains a normal return to low distress after a stressful time. This is the first study to show that a commercially available, facilitator- or self-administered behavioral training product can have significant beneficial effects on psychosocial well-being in a healthy community sample.
Key Words: psychosocial risk factor • stress management • evidence based behavioral medicine • cognitive behavioral therapy • LifeSkills • translational research
Abbreviations: CHD = coronary heart disease; CAD = coronary artery disease; MI = myocardial infarction; CBSM = cognitive–behavioral stress management; WLS = Williams LifeSkills; CABG = coronary artery bypass graft; STAI = State and Trait Anxiety Inventory; CESD = Center for Epidemiological Study Depression scale; ISEL = Interpersonal Support Evaluation List; CM-Ho = Cook-Medley Hostility scale; PSS = Perceived Stress Scale; VO = LifeSkills Video only; WO = LifeSkills Workshop only; WV = LifeSkills Workshop plus LifeSkills Video.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONAPPENDIXNOTESREFERENCES |
|---|
In past studies, behavioral interventions have been successful in both lowering the levels of psychosocial risk factors and improving health outcomes in several populations. Among patients with coronary heart disease (CHD), several studies have implemented cognitive–behavioral interventions to successfully reduce signs of disease (14) such as fewer coronary artery disease (CAD) events (24), reduced diastolic blood pressure (25), lowered resting heart rate (26), and fewer myocardial infarction (MI) recurrences up to 7 years postintervention (27). In addition, as would be expected, many of these interventions simultaneously decrease psychosocial risk factors such as hostility (25), type A behavior (28), stress (27), depression, and social isolation (26). Similar benefits of cognitive–behavioral interventions have also been shown in patients with cancer (29–32).
Although these interventions are effective, their applicability is limited because they often require highly trained professionals to administer them (30–32), the time period over which intervention occurs is lengthy (27,28,32), and/or specialized equipment such as that for biofeedback is required (24). To reach large populations, including those in areas without access to highly specialized resources, more structured, manualized, short-duration interventions are needed.
Extensive research using cognitive–behavioral stress management (CBSM, a structured, manualized group intervention that emphasizes skill learning, cognitive behavioral modification, and relaxation training) with patients with HIV has shown that short-term behavioral therapy is effective at both reducing psychosocial distress and enhancing health (33–40). Structured short-term behavioral interventions like CBSM have also been shown to dramatically decrease medical costs in patient populations (24,41,42). These studies indicate that a manualized, short-term treatment can have long-lasting effects on physical and psychological health in populations with one kind of chronic stress—that associated with diseases like HIV.
Although CBSM and similar interventions have targeted at-risk clinical populations, nonpatients also experience the detrimental health effects of psychosocial risk factors. Advantages of targeting such a healthy, but distressed, community population include its large size and consequent potential for widespread attributable benefit at a public health level as well as the possibility of preventing stress-related diseases such as CHD from developing or progressing. To fill this need, Williams and Williams (43) developed the Williams Life- Skills (WLS) Workshop, a standardized, manualized behavioral intervention designed to be easily administered in community settings.
The WLS Workshop teaches cognitive–behavioral skills and relaxation techniques through six biweekly facilitator-led sessions. In uncontrolled pilot studies, it has been shown to decrease hostility, depression, and social isolation (44). In addition, two randomized, controlled trials have shown that WLS-based interventions decrease hostility and diastolic blood pressure in post-MI patients (25) and decrease depression, anger, and anxiety while improving a variety of cardiovascular outcomes in patients who have undergone coronary artery bypass graft (CABG) surgery (26). Combined, these studies suggest that the WLS workshop can be effective at improving measures of cardiovascular health and at lessening psychosocial risk factors.
Given the limited nature but promising results of past studies, one goal of the current study was to test the WLS Workshop’s short- and long-term effectiveness at reducing psychosocial risk factors known to affect health in a larger, randomized, controlled trial of nonpatients showing moderately elevated psychosocial distress. As a result of the successes of past studies, we predicted that the WLS Workshop would be effective at lessening psychosocial distress over both the short- and long-term.
Analysis of both short- and long-term effectiveness is particularly important in this study because it is the first study to our knowledge to evaluate a CBSM intervention in a nonpatient, healthy sample drawn from the community. In studies such as those of CBSM, in which an illness provides a chronic source of stress, the continued elevated psychosocial distress of untreated control subjects is expected. In this nonpatient sample, however, a continuous source of stress may not be present, perhaps allowing control subjects to improve over time despite lack of treatment. Moreover, even in patient populations, improvement is still sometimes seen in untreated control subjects (45). As a result of this uncertainty about control behavior over time, separate attention to both immediate and long-term effects of training compared with control subjects is needed.
Although the WLS Workshop is easily administered compared with many of the more long-term behavioral treatments, making the therapy self-administered would render it even more accessible to larger community populations. A second goal of this study, then, was to develop a self-administered form of the WLS program.
Self-administration of behavioral interventions has been previously shown to be effective at lessening health-related psychosocial risk factors (46) as has video administration as one specific modality (47). Furthermore, studies using video-based interventions to teach skills unrelated to the WLS material suggest that this medium can have long-lasting effects on behavior and attitudes (48). There is also direct evidence that video is an effective medium for teaching the WLS skill set. In a pilot study of the LifeSkills video, those who were randomized to watching a video segment teaching assertion skills showed improvements in assertion scores and aggression scores as compared with a control group (49). Given the encouraging results of the video pilot and of other studies that used video instruction to alter behavior, we expected a full length LifeSkills Video and accompanying workbook to be more effective than no treatment at lessening psychosocial distress over both short- and long-term follow up. However, we also predicted that the LifeSkills Video would not be as effective as the Workshop, because it lacks the individual attention, supervised practice of skills, and group support that the Workshop provides.
A third goal of the current study was to test the efficacy of a combined LifeSkills Video and Workshop treatment. Although not expected to enhance postintervention well-being over a Workshop-only group, the LifeSkills Video does add an opportunity for continued learning that those in a traditional workshop do not have. Therefore, we predicted that the LifeSkills Video may lead to enhanced maintenance of benefit over a Workshop-only group.
In summary, the aim of the present study was twofold. First, we examined how effective three formats of the Williams LifeSkills program were at reducing several psychosocial risk factors postintervention in a randomized, controlled trial using a community sample showing pretreatment elevations in psychosocial distress. Few studies have been done with any kind of CBSM in such a population, and our initial results from past studies and our own experience administering the system to a variety of clients led us to hypothesize that the Workshop, LifeSkills Video, and LifeSkills Video plus Workshop would all lead to significant improvements over control subjects in the major psychosocial risk factors of anxiety, depression, and perceived stress. Between treatment groups, we hypothesized that the Workshop and LifeSkills Video plus Workshop would be most effective with the LifeSkills Video being slightly less effective because it lacked personalized guidance and group social support.
Our second aim was to investigate maintenance of benefit. We followed participants out to 6 months postintervention to determine whether gains from such a short intervention could be maintained over a long period of time. Given the maintenance of benefit in interventions using similar therapeutic techniques in clinical groups (25,26,38,40), we projected that our system would have long-lasting effects on the psychosocial variables measured. Between treatment groups, we hypothesized that the groups with LifeSkills Videos may show better maintenance than the Workshop only group as a result of participants’ ability to reinforce their skill learning over time by watching the LifeSkills Video again as needed.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONAPPENDIXNOTESREFERENCES |
|---|
Psychosocial Measures
The STAI (form Y) was used to measure trait anxiety (50). To measure depression, the 20-item CESD scale was used because it was designed for use in nonclinical samples such as that of this study (51). Cohen’s Perceived Stress Scale (PSS (9,55)) was used to measure the degree to which participants perceived events in their lives as being stressful. Cohen’s Interpersonal Support Evaluation List (ISEL (52)) was used to evaluate social support and 27 items of the Cook-Medley Hostility scale (CM-Ho (53)) that have been shown to predict mortality in healthy samples was used to measured hostility (54). (See the Appendix.)
Intervention
Participants were randomized to one of four groups: Workshop Only (WO), LifeSkills Video Only (VO), Workshop plus LifeSkills Video (WV) and control (no treatment, CON). All three training conditions consisted of two sessions of training per week for 3 weeks giving six sessions total. Each session lasted between 1 and 2 hours.
Workshop Only
The Williams LifeSkills Workshop (43) is a highly structured psychoeducational group experience that draws on several clinical traditions, including cognitive–behavior therapy (56), relaxation training (57) and skills training. The information is presented in a skills-learning format with the facilitator leading participants through each of several behavioral skills, modeling them as necessary, and allowing time for practice and feedback. The participants were told that the workshop is designed to enable them to cope more effectively with negative emotions and stressful life situations, to improve their relationships with others, and to increase the proportion of positives in their daily lives.
LifeSkills Video Only
The LifeSkills Video and accompanying Workbook are professional quality presentations of all the skills taught in the Workshop. Each skill is presented by actors who model both effective and ineffective skill use followed by a pause for review and practice with the Workbook. The Workbook includes some repetition of particularly important information as well as exercises and homework assignments to encourage practice of the skills.
Although the LifeSkills Video and Workbook sets are designed for independent use at home, to assure compliance, participants in the VO condition were required to come to the Williams LifeSkills Inc. office in Durham, North Carolina, to view LifeSkills Video segments and follow along in the workbook exercises. Participants watched the LifeSkills Video in groups to save time, but they were instructed by a research assistant that no group discussion should occur. At the end of training, all participants received a copy of the LifeSkills Video to take home with them.
LifeSkills Video Plus Workshop
In this condition, participants received the workshop as it was given in the WO group but with use of the LifeSkills Video and workbook integrated into each session. The information received in this training condition was the same as in both VO and WO. All participants received a copy of the LifeSkills Video to take home with them at the end of training.
Control
The control group received no training. After enrollment, randomization, and pretesting, they were instructed to go about their normal lives for the next 3 to 4 weeks after which they returned for another evaluation. This 3- to 4-week delay matched the time period during which treatment groups received training. After completion of the 6-month time point, all controls were given the opportunity to receive the LifeSkills Video and Workbook, but no further data from these participants were included in intervention analyses.
Procedure
All individuals who contacted Williams LifeSkills Inc. were informed that we wanted to study the effects of a behavioral stress management treatment on normal individuals feeling stressed by everyday life. If the individual reported no major illness that met study exclusion criteria, they were invited to come in to the WLS office for a screening session, where informed consent was obtained using a form approved by the Duke University Medical Center Institutional Review Board. In the screening session, potential participants were asked to fill out the six Psychosocial Measures. These measures were scored by a research assistant as they were completed and the potential participant was then informed of whether or not they qualified. Those scoring in the clinical range on the STAI or CESD were referred to a licensed psychologist as described in the "Participants" section. Only once the study psychologist deemed such an individual able to participate did he or she continue as a qualified participant. Potential participants were paid $5 for the screening session regardless of qualification.
Once a group of eight to 10 participants was assembled, a condition was randomly selected (WO, VO, WV, or CON) and all eight to 10 participants were considered enrolled in that group. Each participant was contacted and he or she was asked to return for a preintervention evaluation session. Preintervention evaluation took place approximately 1 to 2 weeks before intervention or the wait period began and generally within a few weeks of when screening had occurred. The delay between screening and the training/wait period was the same for all groups, treatment and control alike. In the preintervention evaluation session, participants were asked to fill out several exploratory questionnaires not included in analysis and several biologic markers (i.e., blood pressure and salivary cortisol) were obtained. The results from the biologic markers will be reported in future publications. Participants were paid $15 for completing the preintervention session. After the successful preevaluation of all participants in a group, the training/waiting period began. Participants in training groups were reimbursed $5 per session to compensate for travel costs to and from the WLS facility.
Participants were evaluated again on all psychosocial measures and biomarkers at three subsequent time points after completion of the training/waiting period: 10 days, 2 months, and 6 months posttraining/wait. These time points were chosen to provide a short-term posttraining assessment, an interim time point that was similar to end points in previous studies (e.g., (25)) and a longer-term assessment than previously used in behavior modification studies. For 10 day and 2-month testing, the participants received $20 each. To maximize the likelihood that participants would return for the last assessment, they received $60 for completing the 6-month time point. Enrollment and randomization of subjects began in January 2002 and 6-month data collection on the last group of subjects was completed in August 2003.
Statistical Analysis
The primary hypothesis tests used mixed-models analysis of variance for repeated measures (PROC MIXED, SAS version 8; SAS Institute, Cary, NC). The mixed-models approach allowed inclusion of incomplete data for subjects who were lost to attrition or missed follow-up visits. The immediate effects of treatment were assessed by analysis of data collected at baseline and 10-day follow up with a test of the treatment group main effects and group by time point interactions. Analysis of long-term follow up included values recorded at baseline, 10 days, 2 months, and 6 months with tests of treatment group main effects and group by time point interactions. When main effects or interactions were significant, treatment groups were separately compared with the control group and with each other to determine where differences were located.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONAPPENDIXNOTESREFERENCES |
|---|
|
|
Baseline Characteristics
There were no significant differences across the four groups at baseline in any of the psychosocial measures. However, mean scores on STAI-T, CESD, and PSS were all above the inclusion cutoff, whereas mean ISEL and CM-Ho scores were within the normal range, indicating that it was increased anxiety, depression, and/or perceived stress that characterized the sample included in this study on average (Fig. 2). Because our recruitment drew a sample in which ISEL and CM-Ho scores were not elevated at baseline, these measures were considered unlikely to show improvement and were not analyzed further.
|
Short-Term Effects of Treatment
Repeated-measures analyses of variance comparing the three treatment and control groups at baseline versus 10 days posttraining/wait period revealed a significant group by time interaction for the STAI-T (F = 4.70[3], p = .0039) and the PSS (F = 2.99[3], p = .0340) but not for CESD. Contrasts of individual treatment groups versus CON revealed better short-term improvement for PSS and STAI-T in the WV (F = 8.43[1], p = .0051 and F = 10.04[1], p = .0024, respectively) and VO (F = 6.28[1], p = .0147 and F = 4.68[1], p = .0343) groups (Fig. 2). The WO group did not differ from CON on any psychosocial factor and there were no significant effects of treatment groups on CESD.
Long-Term Effects of Treatment
Repeated-measures analyses of variance comparing the three treatment and control groups across all four time points revealed a significant group by time interaction for STAI-T (F = 2.33[9], p = .0148) but not CESD or PSS. When treatment groups were compared separately with control, significant group by time point interactions were found for WV versus CON for STAI-T (F = 4.72[3], p = .0034) and for VO versus CON for STAI-T (F = 2.86[3], p = .0385). Although the overall group by time interaction for PSS was not significant, it is of interest to note that the WV versus CON contrast was significant for PSS (F = 2.95[3], p = .0342). Inspection of Figure 2 reveals that these stronger effects of treatment versus CON conditions across the entire follow-up period are primarily the result of the more rapid reduction in scores for the WV and VO groups from pretraining/wait to the 10-day follow up with maintenance of these reductions through 6 months, whereas the CON group does not reach similarly low scores till later in the follow-up period.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONAPPENDIXNOTESREFERENCES |
|---|
Over long-term follow up, both the WV and VO groups showed significant improvement over control subjects on trait anxiety, whereas the WV group also showed significant improvement over control subjects on perceived stress. These findings suggest that the WLS intervention, in particular in the workshop–video combination form, has long-term benefits for psychosocial well-being. However, it should be noted that these effects seem to be mostly the result of the immediate benefits of training, which are maintained as control subjects improve more slowly; by 6 months follow up, all groups have similar levels of trait anxiety and perceived stress. Nonetheless, it is possible that the WLS programs will have longer-term benefits not revealed by this design because participants who received training may retain their skills to better cope with future stressful experiences more effectively than untrained control subjects. Furthermore, there is the possibility that some participants already had the WLS skills set, lessening observed improvement in trained groups.
The finding that the self-administered video format of the WLS program is effective at accelerating the return to normal trait anxiety and perceived stress, as well as maintaining long-term benefits in perceived stress, is particularly important. Although standardized workshops make training available to many more people than individualized training, the delivery of group interventions is still limited by community and individual resources. The self-administered Video format, however, can reach virtually anyone with a VCR or DVD player. Moreover, beyond the normal psychosocially distressed population, the WLS Video could be particularly useful in large populations exposed to major trauma such as acts of terrorism or natural disasters like Hurricane Katrina, which affected millions in the Gulf Coast region. Unlike resource-demanding workshop interventions, the self-administered Video training is sufficiently cost-effective and time-efficient to be administrable to large groups of people, like the Katrina victims, who experience the adverse physical and mental repercussions of a traumatic event.
Although results for trait anxiety and perceived stress were positive, no significant effect of training on depression was found. One possible reason for this finding is that depression may require more intense intervention to be ameliorated than the current training provided. Future studies of the WLS system could investigate this possibility by using more extended versions of the training with some practically administrable individual attention such as through phone calls from trained facilitators during training and beyond.
Like with most studies using a community sample, self-selection bias is a potential problem here. All participants were recruited by community advertisements and required to be able to come to the WLS facilities for evaluations and training. For training involving the workshop format, this issue does not greatly compromise the results because any such intervention in the real world would require travel to a centralized facility. For the video-only group, however, the requirement to come to the WLS facilities means that the video was not completely self-administered. Therefore, the results of this study support the basic principle that merely watching a video and following along with independent exercises in the workbook can reduce psychosocial distress. Further studies with the video and workbook being used by participants in their own homes will be necessary to ensure that benefits are still found when treatment is solely self-administered. Home use of the video will greatly extend it reach to a mass audience.
One important uncertainty in this study was how control scores would change over time. We originally hypothesized that control subjects may improve over follow up as the acutely stressful situations that drew them to the study were resolved. In support of this hypothesis, control subjects were found to slowly improve over the 6-month follow up. Aside from the possibility that this is a normal regression to the mean, it is also important to note that in this study, like in others in which control group improvement has occurred (i.e., the ENRICHD trial (45)), participants were notified that they had elevated level(s) of psychosocial risk factor(s) by their qualification. This notification is an intervention itself, which may have led control subjects to seek other means of reducing their psychosocial distress. In future studies, this problem of control improvement could be minimized through repeated pretraining assessments to ensure distress is maintained before including a participant in randomization.
Nonetheless, the results of this study are notable in that they show that the WLS program can accelerate and maintain improvements in psychosocial well-being over and above improvements seen in control subjects in a healthy but stressed community sample. It is quite possible that the eventual reduction in distress measures in the control group is typical of the "natural history" of transient periods of distress in which a stressful life situation produces an increase in distress that dissipates over time. To the extent that frequent periods of such distress over time contribute to deteriorations in mental and/or physical health, a program like that provided by the WLS Video could shorten one’s exposure to distress and hence improve health. Even in the absence of effects on disease, the shortened periods of distress would still have the important benefit of improving overall quality of life and well-being—a worthy goal in itself.
In summary, the current study shows that the WLS program, a brief, standardized, behavioral stress management intervention, accelerates a return of trait anxiety and perceived stress to normal levels in a healthy but mildly distressed community sample. Moreover, the results also indicate that incorporation of video delivery may contribute to the effectiveness of the training and that having a facilitator-led format in addition to the video may enhance maintenance of benefit because only the WV group achieved significance of posttraining benefits compared with the control group across the entire follow up. Although the kinds of training-related benefits seen here are modest, the ability to reach a mass audience with the video means they could be cumulatively influential at the public health level because they could reduce the average distress level in a large population. However, further research into the longer-term effects of the WLS program and its effects in different populations are needed to clarify the extent of its benefits.
|
APPENDIX |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONAPPENDIXNOTESREFERENCES |
|---|
STAI-S and STAI-T
To meet the distressed inclusion criterion, subjects had to have a score on either State or Trait Anxiety that would place them above community averages their age and gender group (50): men: 19 to 39 years, >38S, >36T; 40 to 49 years, >36S, >35T; 50+ years, >35S, >35T; women: 19 to 39 years, >36S, >36T; 40 to 49 years, >36S, >36T; 50+ years, >33S, >33T. Scores that qualified the participants for a telephone interview with the study psychologist were: men, 19 to 39 years, >57S, >55T; 40 to 49 years, >57S, >54T; 50+ years, >55S, >52T; women, 19 to 39 years, >59S, >56T; 40 to 49 years, >59S, >54T; 50+ years, >50S, >48T.
Center for Epidemiological Study Depression Scale
A score above 10 was considered qualifying, whereas a score above 27 required a phone interview with the study psychologist (51).
Interpersonal Support Evaluation List
A score below 70 was considered qualifying on this measure (52).
Cook-Medley Hostility Scale
Qualifying scores for the CM-Ho were assessed based on age and gender (53): Men: <20 years, >27; 20 to 29 years, >20; 30 to 39 years, >18; 40 to 49 years, >19; 50 to 59 years, >18; 60 to 69 years, >19; 70+ years,>21; women: <20 years, >24; 20 to 29 years, >18; 30 to 39 years, >17; 40 to 49 years, >17; 50 to 59 years, >17; 60 to 69 years, >18; 70+ years, >17.
Perceived Stress Scale
Qualifying scores on the PSS were (8,55): men: 18 to 29 years, >16; 30 to 44 years, >15; 45 to 54 years, >15; 55 to 64 years, >14; 65+ years, >14; women: 18 to 29 years, >18; 30 to 44 years, >16; 45 to 54 years, >16; 55 to 64 years, >15; 65+ years, >15.
|
NOTES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONAPPENDIXNOTESREFERENCES |
|---|
Dr. Redford Williams and Dr. Virginia Williams are founders and majority stock holders in Williams LifeSkills, Inc.
DOI:10.1097/01.psy.0000238452.81926.d3
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONAPPENDIXNOTESREFERENCES |
|---|
This article has been cited by other articles:
 |
|
 |
|
  A. E. Campo, V. Williams, R. B. Williams, M. A. Segundo, D. Lydston, and S. M. Weiss Effects of LifeSkills Training on Medical Students' Performance in Dealing with Complex Clinical Cases Acad Psychiatry, May 1, 2008; 32(3): 188 - 193. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
p < .05 for LifeSkills Workshop plus LifeSkills Video versus control group by time point effect over follow up; ßp < .05 for LifeSkills Video only versus control group by time point effect over follow up; *p < .05 group by time point effect at 10 days; **p < .01 group by time point effect at 10 days; 
p < .05 group by time point interaction.