You are on page 1of 12

Review Article

Guided Imagery for


Arthritis and Other
Rheumatic Diseases: A
Systematic Review of
Randomized Controlled
Trials
--- Peter R. Giacobbi Jr, PhD, Meagan E. Stabler, BS,
Jonathan Stewart, MS, Anna-Marie Jaeschke, MS,
Jean L. Siebert, MA, MBA, and George A. Kelley, DA

- ABSTRACT:
Many individuals suffering from arthritis and other rheumatic diseases
(AORD) supplement pharmacologic treatments with psychosocial in-
terventions. One promising approach, guided imagery, has been re-
ported to have positive results in randomized controlled trials (RCTs)
From the West Virginia University, and is a highly scalable treatment for those with AORD. The main
Morgantown, West Virginia.
purpose of this study was to conduct a systematic review of RCTs that
Address correspondence to Peter R. have examined the effects of guided imagery on pain, function, and
Giacobbi Jr, PhD, Joint Appointment, other outcomes such as anxiety, depression, and quality of life in adults
School of Public Health, West Virginia with AORD. Ten electronic bibliographic databases were searched for
University, Health and Education
Building, 208, P. O. Box 6116, 375 reports of RCTs published between 1960 and 2013. Selection criteria
Birch Street, Morgantown, WV 26506- included adults with AORD who participated in RCTs that used guided
6116. E-mail: peter.giacobbi@mail. imagery as a partial or sole intervention strategy. Risk of bias was as-
wvu.edu
sessed using the Cochrane Risk of Bias Assessment Instrument. Results
Received December 31, 2014; were synthesized qualitatively. Seven studies representing 306
Revised January 28, 2015; enrolled and 287 participants who completed the interventions met
Accepted January 28, 2015.
inclusion criteria. The average age of the participants was 62.9 years
GA Kelley was partially funded by (standard deviation ¼ 12.2). All interventions used guided imagery
the National Institute of General scripts that were delivered via audio technology. The interventions
Medical Sciences of the National In- ranged from a one-time exposure to 16 weeks in duration. Risk of bias
stitutes of Health under award
number U54GM104942. The con- was low or unclear in all but one study. All studies reported
tent is solely the responsibility of the statistically significant improvements in the observed outcomes.
authors and does not necessarily Guided imagery appears to be beneficial for adults with AORD. Future
represent the official views of the
National Institutes of Health.
theory-based studies with cost-benefit analyses are warranted.
Ó 2015 by the American Society for Pain Management Nursing
1524-9042/$36.00
Ó 2015 by the American Society for
Pain Management Nursing
Arthritis and other rheumatic diseases (AORD) are a leading source of disability
http://dx.doi.org/10.1016/ for millions of adults. It is estimated that 55.2 million adults in the United States
j.pmn.2015.01.003 self-report physician-diagnosed arthritis, with estimated prevalence expected to

Pain Management Nursing, Vol 16, No 5 (October), 2015: pp 792-803


Guided Imagery as Treatment for AORD 793

reach more than 67 million by the year 2030 (Barbour 2000; Fors, Sexton, & G€ otestam, 2002; Lewandowski,
et al., 2013). Arthritis and other rheumatic diseases Good, & Draucker, 2005; Menzies, Taylor, &
include more than 100 different conditions that are Bourguignon, 2006). Guided imagery can be defined
typically caused by inflammation, swelling, and pain as a quasi-perceptual, multisensory, and conscious
in patients’ joints, ligaments, bones, muscles, and experience that resembles the actual perception of
sometimes internal organs throughout the body some object, scene, or event but occurs in the absence
(NIAMS, 2014). Adults with AORD often experience of external stimuli (Thomas, 2014). Also known as
declines in lifestyle and recreational physical activity ‘‘mental simulation’’ or ‘‘visualization,’’ this cognitive
and are more prone to depression and anxiety (Covic technique has deep historical roots, scientific interest,
et al., 2012; Kaplan, Huguet, Newsom, & McFarland, and popular applications. Psychologists have long used
2003; Murphy, Sacks, Brady, Hootman, & Chapman, guided imagery to help individuals cope with pain,
2012; Shih, Hootman, Kruger, & Helmick, 2006). anxiety, and trauma (Thomas, 2014). Guided imagery
Although AORD can affect people of all ages, interventions with AORD patients often begin with
rheumatoid arthritis (RA), osteoarthritis (OA), and breathing or progressive muscle relaxation exercises
fibromyalgia are the most common AORD conditions and then proceed to images of movement and physical
experienced by adults, with prevalence estimates in activity free of pain and stiffness (Baird et al., 2010).
the United States of 1.3 million (Helmick et al., Importantly, guided imagery is inexpensive, relatively
2008), 27 million (Lawrence et al., 2008), and 5 million easy to teach, and can be readily applied in both clin-
(Lawrence et al., 2008), respectively. With an ical and community-based settings (Baird et al., 2010;
increasing tendency toward an older population, Giacobbi, Dreisbach, Thurlow, Anand, & Garcia, 2014).
health care costs associated with AORD will likely Given the overlap between the various psychoso-
continue to rise (Hootman & Helmick, 2006), support- cial strategies used to treat AORD (Jensen, 2011), sys-
ing the need for strategies intended to help individuals tematic reviews of one or more of these techniques
cope with chronic pain and augment other treatments helped inform the present review. One team of re-
(Hochberg et al., 2012). searchers systematically reviewed 12 RCTs with partic-
Treatment strategies for AORD generally include a ipants who were diagnosed with fibromyalgia (n ¼ 5),
combination of exercise, diet, and medications osteoarthritis (n ¼ 2), rheumatoid arthritis (n ¼ 1),
(NIAMS, 2014). Body weight management is particu- neck pain (n ¼ 2), pain in the upper limbs (n ¼ 1),
larly important for patients with AORD in order to and diffuse long-term pain (n ¼ 1). The studies
reduce stress on painful joints. Pharmacologic treat- included interventions that used relaxation tech-
ment for AORD depends on the disease being treated niques, massage, biofeedback, the provision of infor-
and the patient’s individual circumstances. For mation, and cognitive behavioral techniques for the
instance, disease-modifying nonbiologic and biologic treatment of musculoskeletal pain (Persson,
medications may be prescribed to patients who have Veenhuizen, Zachrison, & Gard, 2008). Although the
been diagnosed with RA early in the disease (3-6 authors concluded that relaxation training could be
months) without a poor prognosis (Saag et al., 2008). effective at pain reduction, their results should be
Those with RA or OA may also be prescribed nonste- viewed with caution because of the heterogeneity of
roidal anti-inflammatory drugs (MacDonald, 2000), techniques used in the studies reviewed.
whereas only duloxetine, milnacipran, and pregabalin Another systematic review with meta-analysis that
are approved by the Food and Drug Administration for focused on fibromyalgia patients included six RCTs
the treatment of fibromyalgia (NIAMS, 2014). that tested the efficacy of hypnosis and guided imagery
Because of the side effects, risks, financial bur- on pain, sleep, fatigue, depressed mood, and health-
dens, and patient dissatisfaction with common phar- related quality of life (Bernardy, Fuber, Klose, &
macologic treatments (Nestoriuc, Orav, Liang, Horne, Hauser, 2011). The authors of this review included
& Barsky, 2010; Page & Henry, 2000; Taylor, Everett, studies that combined imagery and hypnosis and
Taylor, Watson, & Taylor-Stokes, 2013; Woolf et al., pointed out that both approaches attempt to promote
2004), many individuals suffering from AORD resort changes in emotion, sensation, perception, thought,
to psychosocial strategies. These may include, but and behavior by offering suggestion. Although meta-
are not limited to, relaxation, mindfulness analytic results indicated significant reductions in
meditation, or hypnosis (Jensen, 2011). Guided imag- pain, these observations were tempered by low meth-
ery has been reported to have positive results, with odologic quality of the studies reviewed (adequacy of
respect to AORD-related outcomes, in randomized randomization, blinding of outcome assessor, and
controlled trials (RCTs) (Baird, Murawski, & Wu, lack of intent-to-treat analyses). Effect sizes on the
2010; Baird & Sands, 2004, 2006; Fors & Gotestamm, other outcomes considered in this meta-analysis were
794 Giacobbi Jr et al.

not calculated because of limited data. Collectively, the included ‘‘random,’’ ‘‘mental imagery,’’ ‘‘guided imagery,’’
mixed results from these previous systematic reviews ‘‘visualization,’’ and ‘‘relaxation’’; also used were
and meta-analyses demonstrate the need for more care- ‘‘randomly,’’ ‘‘randomized,’’ and ‘‘randomized’’ to in-
ful characterization of the psychosocial interventions crease possible retrieval. The fifth author, a Health Sci-
included by previous authors. ences librarian, conducted all searches in consultation
The purpose of this study was to conduct a sys- with the research team. An example of the search strat-
tematic review of RCTs that have examined the effects egy used for one of the electronic databases (Medline
of guided imagery in adults with AORD in order to from Ebscohost) is available upon request to the corre-
determine whether this intervention approach is effec- sponding author. No attempt was made to search for
tive at reducing pain, increasing function, or unpublished data because previous research has sug-
improving other outcomes such as anxiety, depression, gested that these efforts may not be worthwhile (van
and quality of life. A secondary purpose was to charac- Driel, de Sutter, Maeseneer, & Christiaens, 2009).
terize the theoretical underpinnings and nature of im-
agery exposure by participants in RCTs used to treat
Study Selection
AORD outcomes. The present study extends previous
Studies were selected by the first three authors who
systematic reviews on the use of psychosocial strate-
independently reviewed all studies in close consulta-
gies for the treatment of AORD by focusing on studies
tion with study personnel. During and between peri-
that used guided imagery and coding key methodo-
odic meetings, the studies were reviewed for
logic information not addressed in previous systematic
accuracy and consistency. If consensus could not be
reviews (e.g., theoretical frameworks, length of inter-
achieved, the last author was consulted and asked for
vention, and cost-benefit analyses). The coded studies
a recommendation. A list of included and excluded
were not meta-analyzed because of methodologic het-
studies, incorporating reasons for exclusion, was
erogeneity between studies.
stored in a Microsoft Excel 2013 file.

METHODS Coding Sheet and Data Extraction


Study Eligibility Criteria The codebooks were developed by the first author
Inclusion criteria included the following: (1) RCTs working closely with the senior investigator (GAK)
with a comparison group; (2) adult participants aged on the team. The major categories of variables coded
18 years and older; (3) use of guided imagery as the included the following: (1) study characteristics (year
sole or partial intervention strategy; (4) focus on of publication, journal); (2) participant demographics;
AORD; (5) publications in English from January 1, (3) length of intervention; (4) mode of intervention de-
1960 to June 1, 2013; and (6) results reported for livery; and (5) primary and secondary outcomes
pain, physical function, anxiety, depression, or quality measured. Three doctoral students (MS, JS, AMJ) inde-
of life. Studies were limited to RCTs because this pendently coded studies that met the inclusion
research approach is the only way to control, by study criteria. Risk of bias was evaluated using the Cochrane
design, for known confounders and the observation risk of bias tool and included six known sources of bias
that nonrandomized approaches tend to overestimate in RCTs (Higgins & Green, 2009): (1) blinding of study
treatment effects (Sacks, Chalmers, & Smith, 1982; personnel or participants to group assignment; (2)
Schulz, Chalmers, Hayes, & Altman, 1995). sequence generation of group assignment; (3) alloca-
tion of participants to treatment groups; (4) incom-
Data Sources plete outcome data; (5) incomplete outcome
Citations were retrieved from searching 10 electronic reporting; (6) and other sources of bias (Higgins &
bibliographic databases (Academic Search Complete, Green, 2009). An important part of the coding process
Medline from Ebscohost, PsycInfo, Scopus, SPORTDis- was to observe the length, nature, timing, and mode of
cus, Cochrane Central Register of Controlled Clinical intervention delivery for the reviewed studies. This de-
Trials, Cumulative Index to Nursing and Allied Health cision was based on the flexibility for intervention de-
Literature, Physiotherapy Evidence Database, Web of livery with guided imagery because this treatment
Science, and ERIC), (2) cross-referencing from approach can be delivered in person, by paper, or us-
retrieved studies, including systematic reviews and ing electronic methods. Likewise, studies were exam-
meta-analyses, and (3) hand searching specific jour- ined regarding whether cost-benefit analyses were
nals. Variations of specific keywords were tested for conducted. Finally, studies were coded on a priori
relevancy to our topic and whether truncation would theoretical frameworks and if previous theorizing
work best in different databases. Searched keywords informed the content of imagery exposure.
Guided Imagery as Treatment for AORD 795

Data Synthesis of pain (Lewandowski, 2004), fibromyalgia impact


Because of the expected heterogeneity with respect to questionnaire, arthritis self-efficacy (Menzies et al.,
such things as study design, participant characteristics, 2006), and journaling to measure medication usage
intervention, and outcome variables being measured, (Baird et al., 2010).
an a priori decision was made to assess all results Statistically significant results supporting the use
qualitatively. of guided imagery were observed in all seven studies.
In Baird et al. (2010), statistically significant reductions
in pain and medication usage were accompanied by
RESULTS increased function and mobility with their 16-week
The characteristics of the studies reviewed are listed in intervention. Results by Fors et al. (2002) and Fors
Table 1. Of the 1,313 studies reviewed, 7 met the inclu- and Gotestamm (2000) revealed acute reductions for
sion criteria. The studies included 16 groups (9 inter- pain and anxiety in a laboratory setting immediately af-
vention and 7 control) representing 306 individuals, ter exposure to imagery, whereas their efforts in 2002
with 8 men and 282 women randomly assigned to reflected daily exposure to imagery and reductions in
the various study arms (Baird et al., 2010; Baird & pain but not anxiety. Lewandowski’s (2004) 4-day
Sands, 2004, 2006; Fors et al., 2002; Fors & study revealed that participants exposed to guided im-
Gotestamm, 2000; Lewandowski, 2004; Menzies agery reported significantly reduced pain, compared
et al., 2006). The selection process and reasons for with the control group, during the last 2 days of the
exclusion are shown in Figure 1, and a full list of study. The only study to use a disease-specific measure-
excluded studies is available upon request to the corre- ment scale produced statistically significant reductions
sponding author. in fibromyalgia symptoms and coping self-efficacy re-
sulting from listening to a 20-minute audio file once
daily for 21 days (Menzies et al., 2006). Finally, none
Imagery Exposure
of the seven studies performed a cost-benefit analysis.
Imagery exposure was compared with control condi-
tions and was described by the authors as usual care
Intervention Characteristics and Theoretical
plus journaling (Baird and Sands, 2004, 2006), a
Underpinnings
sham intervention involving planned rest, daily logs
Baird and Sands (2004, 2006) combined guided
of medication use, weekly pain ratings (Baird et al.,
imagery with progressive muscle relaxation, whereas
2010), pain-related talk with a therapist (Fors and
other investigators combined different variations of
Gotestamm, 2000), or usual care (Fors et al., 2002;
relaxation exercises within their imagery
Lewandowski, 2004; Menzies et al., 2006).
interventions (Baird et al., 2010; Fors et al., 2002;
Lewandowski, 2004; Menzies et al., 2006). Only one
Participant Characteristics study did not report the use of relaxation exercises
The sample sizes of the seven studies ranged from 28 embedded within the imagery intervention (Fors &
to 58 participants with an average age of 62.9 years. Gotestamm, 2000).
The breakdown by gender was disproportionate All seven interventions involved scripts delivered
because only two of the seven studies included men with audio files. Researchers in one of the seven
(Lewandowski, 2004; Menzies et al., 2006). One studies used a combination of researcher instructions
study included contradictory reports of gender and audio delivery of the imagery scripts (Fors &
breakdown in the published manuscript (Baird et al., Gotestamm, 2000). Fors and Gotestamm (2000)
2010). required participants to visit a clinical setting for
administration of their intervention with sessions last-
Outcomes ing 8 weeks. The remaining six studies involved
All seven studies relied on self-report surveys to mea- home-based imagery practice (Baird et al., 2010;
sure the primary and secondary outcomes shown in Baird & Sands, 2004, 2006; Fors et al., 2002;
Table 2. These outcomes included pain in four of the Lewandowski, 2004; Menzies et al., 2006).
studies (Baird et al., 2010; Fors & Gotestamm, 2000; Between pre- and post-testing, the imagery inter-
Lewandowski, 2004; Menzies et al., 2006), ventions lasted 30 minutes (Fors & Gotestamm,
psychological well-being in three studies (Baird et al., 2000) to 16 weeks (Baird et al., 2010). In a pair of
2010; Baird & Sands, 2004, 2006), and anxiety and 12-week investigations, participants were instructed
depression in two (Fors et al., 2002; Fors & to listen to their imagery audio files twice daily but
Gotestamm, 2000) and one study, respectively (Fors no information was provided about the length of the
et al., 2002). Other outcomes included the meaning audio files (Baird & Sands, 2004, 2006). It is
796 Giacobbi Jr et al.

TABLE 1.
Study Characteristics and Theoretical Bases
Theoretical
Study Country Participants Imagery Intervention Underpinnings

Baird & Sands, United States 28 women aged 65 years 12 weeks listening to Psychoneuromuscular
2004 and older diagnosed audio files twice daily theory
with OA were assigned
to GI plus PMR
(Mage ¼ 72.1) or usual
care (Mage ¼ 74.8)
Baird & Sands, United States Same sample as in Baird & 12 weeks listening to Biopsychosocial and
2006 Sands, 2004 audio files twice daily psychoneuromuscular
theories
Baird et al., 2010United States 30 older adults, gender 16 weeks listening to Biopsychosocial and
breakdown unclear, audio files (12 minutes) psychoneuromuscular
who self-reported OA, twice daily theories
moderate to severe
pain, and mobility
difficulties were
randomized to GI plus
relaxation (Mage ¼ 71.9)
or a sham intervention
(Mage 62.6) (planned
rest)
Fors & Gotestamm, Norway 58 women (Mage 45.7) 30-minute ‘‘audio Not discussed
2000 diagnosed with instructions’’
fibromyalgia were
randomized to patient
education (n ¼ 22),
guided imagery (n ¼ 17),
or a pain-related talk
group (n ¼ 19)
Fors et al., 2002 Norway Same Fors & Gotestamm, 4 weeks; audio files Hypervigilance theory;
2000 with 3 dropouts adaptation theory
in the latter group
Lewandowski, 2004 United States 37 women and 7 men 7-minute audio tape Science of ‘‘unitary beings’’
between 34-90 years of three times a day
age (median age ¼ 61) for 4 days
who self-reported pain
diagnoses from arthritis,
fibromyalgia, and other
conditions were
randomized to guided
imagery (n ¼ 21) or
control groups (n ¼ 21)
Menzies et al., United States 47 women and 1 man 6 weeks; audio files Not discussed
2006 (Mage ¼ 49.6) were
randomized GI or usual
care
GI ¼ guided imagery; PMR ¼ progressive muscle relaxation; OA ¼ osteoarthritis.

important to note that Baird and Sands (2004, 2006) of imagery exposure to a 30-minute audio file adminis-
relied on the same sample but reported outcomes tered in a lab setting as previously described (Fors &
related to self-reported pain and physical function in Gotestamm, 2000): This sample of participants and
the earlier manuscript and psychological well-being group assignments were then part of a prospective 4-
in the later manuscript. Similarly, Fors and week home-based intervention using the same audio
Gotestamm (2000) and Fors et al. (2002) used the files, which consisted of once daily practice (Fors
same sample with one focusing on the acute effects et al., 2002). A 16-week investigation also used twice
Guided Imagery as Treatment for AORD 797

FIGURE 1. - Flow diagram for selection of studies.

daily practice with an imagery audio file that lasted 12 Theoretical underpinnings of the imagery interven-
minutes (Baird et al., 2010). Lewandowski (2004) and tions were diverse and ranged from no discussion of theo-
Menzies et al., 2006 lasted 7 weeks using a 7-minute retical bases (Fors & Gotestamm, 2000; Menzies et al.,
audio file practiced three times daily (Lewandowski, 2006), an ambiguous presentation in one study
2004), and 6 weeks using an audio file lasting 20 mi- (Lewandowski, 2004), to somewhat more sophisticated
nutes practiced once daily, respectively (Menzies theorizing in others (Baird et al., 2010; Baird & Sands,
et al., 2006). 2004, 2006). Psychoneuromuscular (Jacobsen, 1932)
The nature and content of the imagery scripts and biopsychosocial (Engel, 1977) theoretical models
and authors’ reporting of information varied between were used in one series of studies with the same lead
studies. One study developed personalized scripts author (Baird et al., 2010; Baird & Sands, 2004, 2006,).
that were based on interviews of participants and Briefly, psychoneuromuscular theory predicts that
included descriptions of specific joints and move- guided imagery may stimulate neurologic pathways
ments that caused pain and being in a relaxing place between the motor cortex of the brain to implicated
(Baird & Sands, 2004). Two other studies lead by musculoskeletal systems in a similar manner, albeit a
Baird and Sands (2006) and Baird et al. (2010) also lower amplitude, to when the movements are actually
co-developed imagery scripts with participants that performed. The links between the guided imagery
were characterized by attempts to focus participants’ scripts used by Baird et al. (2010) and Baird and Sands
attention to imagine physical movements without (2004, 2006), with psychoneuromuscular theory
stiffness, pain, or hesitancy. The remaining four focused on the specific movements that caused pain
studies developed universal imagery scripts whereby rather than the pain itself. Biopsychosocial theory
all research participants were exposed to the same (Engel, 1977) was the framework for addressing sensa-
imagery scripts in the intervention (Fors et al., tions of pain by Baird et al. (2010) and Baird and Sands
2002; Fors & Gotestamm, 2000; Lewandowski, (2004, 2006). This theory predicts that multiple factors
2004; Menzies et al., 2006). Only one study affect pain sensation and that imagery may initiate
mentioned embedding music within the audio file adaptive cognitive processes such as active coping,
(Fors et al., 2002). No information was provided refocusing attention, distraction (Schoenfeld-Smith
about the voice characteristics within audio files in et al., 1996), reduced autonomic responses, reduced
any of the studies reviewed. muscle contractions, and other responses similar to
798 Giacobbi Jr et al.

TABLE 2.
Outcome Measures and Authors’ Conclusions
Study Outcome Measures Authors’ Conclusions

Baird & Sands, 2004 AIMS-2; mobility survey not reported Significant reductions in pain and mobility
difficulties at week 12 compared
with control group
Baird & Sands, 2006 Health related quality of life (AIMS-2) Significant increases in health-related quality
of life at week 12 compared with
control group
Baird et al., 2010 Numeric rating scales of pain; mobility Significant reductions in pain compared with
difficulty (AIMS-2 [short forms]); the control group after 4 months;
Western McMaster Osteoarthritis Significant improvements in
Scale; journals to measure mobility from baseline to month 2;
medication use Significant reductions in
prescribed analgesics from
baseline to month 4 and total
medication use from baseline to
month 2 and months 2 to 4
Fors & Gotestamm, 2000 Pain and anxiety using a VASs Patient education and guided imagery groups
experienced significant reductions in pain
and anxiety
Fors et al., 2002 Anxiety (STAI-T); depression (BDI); and Significant reductions in pain for those in the
automatic negative thoughts pleasant imagery condition compared with
(ATQ-30) control group; no difference between
attention imagery and control group
Lewandowski, 2004 Pain using a VAS and the McGill Pain Perceived pain improved in the treatment
Questionnaire; Power as Knowing group
Participation in Change Tool; Imagery
Ability Questionnaire; Marlow-Crowne
Social Desirability Scale
Menzies et al., 2006 McGill Pain Questionnaire (short form); Significant reductions in FIQ scores in group
impact scores; Fibromyalgia Impact exposed to imagery compared to controls;
Questionnaire (FIQ); Arthritis Self-Efficacy self-efficacy for managing pain and other
Scale symptoms of FM increased significantly
with intervention group compared to
controls
AIMS-2 ¼ Arthritis Impact Measurement Scale–2; VAS ¼ visual analog scale; STAI-T ¼ State-Trait Anxiety Inventory–Trait; BDI ¼ Beck Depression Inventory;
ATQ-30 ¼ Automatic Negative Thoughts Inventory; FM ¼ fibromyalgia.

those occurring when individuals experience stress. An perspectives were experimentally manipulated by
examination of ancillary material provided by Baird and Fors et al. (2002).
Sands (2004) showed clear connections between the im- Finally, science of unitary beings (Rogers, 1992)
agery scripts and their theoretical frameworks. was used as a conceptual model in one study
Hypervigilance and adaptation theories informed (Lewandowski, 2004). This viewpoint holds that the
theorizing in one study representing approach versus experience of chronic pain is a lifestyle that evolves
avoidance imagery (Fors et al., 2002). The former the- as a dynamic interplay between the person and envi-
ory suggests that pain would increase with time when ronment and maintaining power, relationships, har-
an individual focuses on the experience of pain and mony, and control are ways of adapting and evolving
that chronic pain sufferers are overly focused on pain as a person with pain. The imagery scripts used in
(Chapman, 1986). In contrast, adaptation theory sug- this investigation targeted relaxation, sensory images
gests that pain would diminish when an individual related to pain, and sensory images intended to create
repeatedly focuses on pain perceptions (Naliboff, personal change.
Cohen, Schandler, & Heinrich, 1981). Hypervigilance
and adaptation are more accurately viewed as hypoth- Risk of Bias
eses or contrasting perspectives on the pain experi- Risk of bias results is shown in Table 3 and Figure 2. As
ence and not theories. These contrasting pain can be seen, a large percentage of items across all
Guided Imagery as Treatment for AORD 799

TABLE 3.
Cochrane Collaboration Risk of Bias
Baird & Baird & Fors &
Sands, Sands, Baird et al., Gotestamm, Fors et al., Lewandowski Menzies
Risk of Bias 2004 2006 2010 2000 2002 et al., 2005 et al., 2006

Sequence Unclear Low Unclear Unclear Low Low Low


Allocation Unclear Unclear Unclear Unclear Low Unclear Low
Blinding Unclear Unclear Unclear Unclear Low Unclear Unclear
Incomplete Low Low Low High Low Low Low
outcome data
Incomplete Unclear Low Low Low Low Low Low
outcome reporting

categories were classified as being at an unclear or low were classified as being at a low risk of bias (Baird
risk of bias. For sequence generation, four studies et al., 2010; Baird & Sands, 2004, 2006; Fors et al.,
were at a low risk of bias (Baird & Sands, 2006; Fors 2002; Lewandowski, 2004; Menzies et al., 2006) and
et al., 2002; Lewandowski, 2004; Menzies et al., one was considered to be at a high risk (Fors &
2006), whereas the remaining three were at an Gotestamm, 2000). Finally, six studies were coded as
unclear risk of bias (Baird & Sands, 2004, Baird et al., low risk of bias for incomplete outcome reporting
2010; Fors & Gotestamm, 2000). For allocation (Baird et al., 2010; Baird & Sands, 2006; Fors et al.,
concealment, two studies were considered to be at 2002; Fors & Gotestamm, 2000; Lewandowski et al.,
low risk of bias (Fors et al., 2002; Menzies et al., 2005; Menzies et al., 2006) with one being unclear
2006), whereas five were considered to be at an (Baird & Sands, 2004).
unclear risk (Baird et al., 2010; Baird & Sands, 2004, Within each study, one was categorized as low risk
2006; Fors & Gotestamm, 2000; Lewandowski, across all risk of bias domains (Fors et al., 2002), one
2004). Six studies were classified as being at an was classified as low risk across four domains
unclear risk of bias for blinding (Baird et al., 2010; (Menzies et al., 2006), two were classified as low risk
Baird & Sands, 2004, 2006; Fors & Gotestamm, 2000; across two domains (Baird et al., 2010; Baird &
Lewandowski et al., 2005; Menzies et al., 2006) and Sands, 2004), and two were classified as low risk
one was considered low risk (Fors & Gotestamm, across three domains (Baird & Sands, 2006;
2000). For incomplete outcome data, six studies Lewandowski, 2004). One study was considered to

FIGURE 2. - Pooled Cochrane Collaboration risk of bias.


800 Giacobbi Jr et al.

be at a high risk of bias in one domain (Fors & RCTs that test the impact of guided imagery on
Gotestamm, 2000). With the exception of one study AORD and may include assessments of strength, gait,
(Fors et al., 2002), all others included at least one balance, endurance, and other functional outcomes
domain that was classified as being at an unclear risk that can be readily administered in both clinical and
of bias (Baird et al., 2010; Baird & Sands, 2004, 2006; community-based settings. These variables should be
Fors & Gotestamm, 2000; Lewandowski et al., 2005; measured while blinding study personnel to group
Menzies et al., 2006). assignment of participants. Future investigators should
also assess participant compliance with assigned imag-
ery exposures in order to measure dose-response rela-
DISCUSSION tionships and evaluate potential causal links between
Overall Findings guided imagery exposure and outcomes. Although a
The current systematic review provides evidence, with variety of procedures are possible, investigators could
certain qualifiers, that guided imagery is an effective administer part or all of the imagery intervention in a
intervention for the treatment of AORD-related health laboratory setting where participants are administered
conditions. Specifically, all seven studies reported re- audio scripts and time of exposure is measured.
sults that support the use of guided imagery as a ther- These recommendations highlight qualifiers and
apeutic tool for the treatment of pain (Baird et al., shortcomings of the RCTs in the present review. First,
2010; Baird & Sands, 2004; Fors et al., 2002; Fors & there was an unclear risk of bias for several of the
Gotestamm, 2000; Lewandowski, 2004), improved studies with respect to study design (e.g., blinding,
psychological well-being (Baird & Sands, 2006; sequence generation, allocation concealment) (Baird
Menzies et al., 2006), improved mobility (Baird et al., et al., 2010; Baird & Sands, 2004; Fors & Gotestamm,
2010; Baird & Sands, 2004), reduced anxiety (Fors & 2000). These concerns should be addressed in future
Gotestamm, 2000), and increased self-efficacy manag- RCTs with clear reporting of study design information
ing pain and fibromyalgia symptoms (Menzies et al., consistent with Consolidated Standards of Reporting
2006). The reviewed studies were relatively short, pre- Trials (CONSORT) guidelines (Moher et al., 2010). A
sumably inexpensive, and administered in either a second shortcoming of the reviewed studies was that
clinic or home-based setting. Given the overall find- two sets of studies used the same sample that was re-
ings, these results suggest that guided imagery is ported in separate publications. Baird and Sands
worthy of further investigation given its observed ben- (2004) reported improvements in pain and mobility,
efits and potential cost-effectiveness in the treatment whereas improved psychological well-being was re-
of adults with AORD. ported in the second publication (Baird & Sands,
2006). Fors and Gotestamm (2000) and Fors et al.
Implications for Research (2002) also reported data from the same randomized
Several implications for future research are gleaned participants in separate publications; however, their
from this systematic review. First, there is a need for second study involved a longitudinal comparative anal-
larger and longer RCTs that examine the impact of ysis with participants originally allocated to the three
guided imagery on AORD outcomes. Such trials could treatment arms. As discussed earlier, a third short-
be creatively administered using various technology coming of the reviewed studies was reliance on
platforms that include the internet, telephone, or the self-report for all the major outcomes, which raises
use of prerecorded compact discs (CD). Indeed, these the possibility of response biases.
technologies create possibilities of population-based As discussed in this review, many of the RCTs used
trials that should include more male participants and in- guided imagery along with relaxation exercises. From a
dividuals from diverse sociodemographic backgrounds. scientific standpoint, imagery interventions should not
These trials should explore potential dose-response ef- include relaxation techniques so that the independent
fects and include follow-up measures to evaluate the impact of guided imagery can be evaluated. Despite
impact of continued guided imagery practice over this scientific ideal, researchers typically combine im-
longer periods. In addition, to effectively implement agery with relaxation because of common cognitive
GI treatment, the feasibility and acceptability of GI processes between psychosocial techniques for the
practices within populations of interest need to be treatment of pain (Jensen, 2011). Therefore, an impor-
well established via empirical future studies. tant question for this review is how guided imagery is
Greater methodologic rigor should also be consid- distinct from, but also complementary to, mindfulness
ered particularly with regard to the use of outcome meditation (MM), hypnosis, and relaxation techniques
measures and dose-response relationships. For for the treatment of pain related to AORD. Guided im-
instance, objective measures could be used in future agery, like mindfulness meditation, hypnosis, and
Guided Imagery as Treatment for AORD 801

relaxation techniques, directs cognitive attention and to use guided imagery scripts with demonstrated effec-
increases awareness of individuals’ thoughts, feelings, tiveness in research studies by creating audio scripts
and behaviors through acceptance of the pain and with documented effect. Standardized scripts could
self-efficacy building visualizations (Jensen, 2011). also be edited and personalized for individual patients.
Each of these techniques involves deliberate practice
and requires conscious efforts to focus one’s mind on
Potential Limitations of the Current Study
breathing and or visualization. Autonomy is a central
The current study is subject to several potential limita-
theme in these techniques because one could presum-
tions. For example, only published studies were
ably choose to engage in guided imagery, relaxation, or
included in the review process. Thus, the results may
mindfulness at any time or place: Hypnosis requires in-
have been influenced by publication bias because
duction and likely cannot be practiced in as many set-
studies that did not reach publication were not
tings as the other techniques. Each of these techniques
included. It is also possible that other inclusion and
can also be self-taught or instructor guided. Guided im-
exclusion criteria may have led to selection bias. These
agery training typically involves prewritten scripts,
include language (only those published in English), age
prompts, or outlines to facilitate individualized medita-
limit (only studies with participants aged 18 years and
tion and imagery. All these methods can be safely and
older), and the limitations of the time frame for publi-
inexpensively delivered to various populations as a
cation. Many of the studies included in the review
supplement to medication to provide comfort, relaxa-
used self-report questionnaires to collect data. As a
tion, and a means of coping with pain and adversity.
result, the findings of the current study are also subject
to errors in self-report questionnaires such as self-
Implications for Nursing Practice
report bias. The generalizability of these findings might
The results of this review provide justification for
also be limited based on the participants included in
nursing professionals to use guided imagery in clinical
each of the studies. Finally, although the researchers
settings. Nursing personnel could facilitate the use of
followed the Cochrane Collaboration guidelines for
guided imagery with television, CDs, and the internet.
coding risk of bias (Higgins & Green, 2009), there is
Given the close association noted by other authors
still a substantial amount of subjectivity in these assess-
(Giacobbi et al., 2014) regarding links between verbal-
ments. As a result, others may have made different de-
izations of experience and guided imagery, nurses
cisions in how selected items should be assessed.
could ask patients to describe various physical activ-
ities and movements that are conducted free of pain
and stiffness. If patients respond with great detail to
these questions, it is likely they are engaging in mental
CONCLUSIONS
imagery. Time permitting, nursing or other health The results of this qualitative systematic review
personnel could co-develop guided imagery scripts suggest that guided imagery may improve selected out-
with patients in order to maximize the impact of imag- comes in adults with AORD. Additional, well-designed
ery on patient outcomes. Self-efficacy building state- randomized controlled trials are needed to more fully
ments should be embedded into the scripts. If substantiate AORD outcomes from the use of guided
possible, clinical personnel could maintain libraries of imagery. However, practitioners are encouraged to
guided imagery scripts that can be used, shared, and implement guided imagery in clinical settings using
reused over time. A more efficient process would be various technologies.

REFERENCES
Baird, C. L., Murawski, M. M., & Wu, J. (2010). Efficacy of women with osteoarthritis. Research in Nursing & Health,
guided imagery with relaxation for osteoarthritis symptoms 29(5), 442–451.
and medication intake. Pain Management Nursing: Official Barbour, K. E., Stevens, J. A., Helmick, C. G., Luo, Y. H.,
Journal of the American Society of Pain Management Murphy, L. B., Hootman, J. M., Theis, K., Anderson, L. A.,
Nurses, 11(1), 56–65. Baker, N. A., & Sugerman, D. E. (2013). Prevalence of doctor-
Baird, C. L., & Sands, L. (2004). A pilot study of the diagnosed arthritis and arthritis-attributable activity
effectiveness of guided imagery with progressive muscle limitation-United States. MMWR Morbidity and Mortality
relaxation to reduce chronic pain and mobility difficulties Weekly Report, 62(44), 869–873.
of osteoarthritis. Pain Management Nursing, 5(3), Bernardy, K., Fuber, N., Klose, P., & Hauser, W. (2011).
97–104. Efficacy of hypnosis/guided imagery in fibromyalgia syn-
Baird, C. L., & Sands, L. P. (2006). Effect of guided imagery drome—a systematic review and meta-analysis of controlled
with relaxation on health-related quality of life in older trials. BMC Musculoskelet Disordors, 12, 133.
802 Giacobbi Jr et al.

Chapman, C. (Ed.). (1986). Pain, perception, and illusion. MacDonald, T. M. (2000). Epidemiology and pharmacoe-
New York: Raven Press. conomic implications of non-steroidal anti-inflammatory
Covic, T., Cumming, S. R., Pallant, J. F., Manolios, N., drug-associated gastrointestinal toxicity. Rheumatology,
Emery, P., Conaghan, P. G., & Tennant, A. (2012). Depression 39(Suppl 2), 13–20.
and anxiety in patients with rheumatoid arthritis: prevalence Menzies, V., Taylor, A. G., & Bourguignon, C. (2006). Ef-
rates based on a comparison of the Depression, Anxiety and fects of guided imagery on outcomes of pain, functional
Stress Scale (DASS) and the hospital, Anxiety and Depression status, and self-efficacy in persons diagnosed with fibromy-
Scale (HADS). BMC Psychiatry, 12, 6. algia. The Journal of Alternative and Complementary
Engel, G. L. (1977). The need for a new medical model: A Medicine, 12(1), 23–30.
challenge for biomedicine. Science, 196(4286), 129–136. Moher, D., Hopewell, S., Schulz, K. F., Montori, V.,
Fors, E. A., & Gotestamm, K. G. (2000). Patient education, Gotzsche, P. C., Devereaux, P. J., Elbourne, D., Egger, M.,
guided imagery and pain related talk in fibromyalgia coping. & Altman, D. G. (2010). CONSORT 2010 Explanation
European Journal of Psychiatry, 14(4), 233–240. and Elaboration: Updated guidelines for reporting
Fors, E. A., Sexton, H., & G€ otestam, K. G. (2002). The ef- parallel group randomised trials. J Clin Epidemiol, 63(8),
fect of guided imagery and amitriptyline on daily fibromyal- e1–37.
gia pain: a prospective, randomized, controlled trial. Journal Murphy, L. B., Sacks, J. J., Brady, T. J., Hootman, J. M., &
of Psychiatric Research, 36(3), 179–187. Chapman, D. P. (2012). Anxiety and depression among US
Giacobbi P. R. Jr., Dreisbach, K. A., Thurlow, N. M., adults with arthritis: Prevalence and correlates. Arthritis
Anand, P., & Garcia, F. (2014). Mental imagery increases self- Care & Research, 64(7), 968–976.
determined motivation to exercise with university enrolled Naliboff, B., Cohen, M., Schandler, S., & Heinrich, R.
women: A randomized controlled trial. Psychology of Sport (1981). Signal detection for chronic back patients and cohort
and Exercise, 15, 374–381. controls to radiant heat stimuli. Journal of Abnormal Psy-
Helmick, C. G., Felson, D. T., Lawrence, R. C., Gabriel, S., chology, 90, 271–274.
Hirsch, R., Kwoh, C. K., Liang, M. H., Kremers, H. M., Nestoriuc, Y., Orav, E. J., Liang, M. H., Horne, R., &
Mayes, M. D., Merkel, P. A., Pillemer, S. R., Reveille, J. D., Barsky, A. J. (2010). Beliefs about medicines predict non-
Stone, J. H., & National Arthritis Data Workgroup (2008). specific side effects in rheumatoid arthritis patients.
Estimates of the prevalence of arthritis and other rheumatic Arthritis Care & Research, 62(6), 791–799.
conditions in the United States: Part I. Arthritis & Rheuma- NIAMS (2014). Arthritis and rheumatic diseases. Be-
tism, 58(1), 15–25. thesda, MD: National Institutes of Health. Retrieved January
Higgins, J. P. T., & Green, S. (2009). Cochrane handbooks 15, 2015 from. http://www.niams.nih.gov/Health_Info/
for reviews of interventions. Chichester, West Sussex, En- Arthritis/arthritis_rheumatic_qa.asp#3.
gland: The Cochrane Collaboration and John Wiley & Sons, Page, J., & Henry, D. (2000). Consumption of NSAIDs and
Ltd. the development of congestive heart failure in elderly pa-
Hochberg, M. C., Altman, R. D., April, K. T., Benkhalti, M., tients. Archives of Internal Medicine, 160, 774–784.
Guyatt, G., McGowan, J., Towheed, T., Welch, V., Wells, G., Persson, A. L., Veenhuizen, H., Zachrison, L., & Gard, G.
Tugwell, P., & American College of Rheumatology (2012). (2008). Relaxation as treatment for chronic musculoskeletal
American College of Rheumatology 2012 recommendations pain—a systematic review of randomised controlled studies.
for the use of nonpharmacologic and pharmacologic thera- Physical Therapy Reviews, 13(5), 355–365.
pies in osteoarthritis of the hand, hip, and knee. Arthritis Rogers, M. E. (1992). Nursing science and the space age.
Care & Research, 64(4), 465–474. Nursing Science Quarterly, 5, 27–34.
Hootman, J. M., & Helmick, C. G. (2006). Projections of US Saag, K. G., Teng, G. G., Patkar, N. M., Anuntiyo, J.,
prevalence of arthritis and associated activity limitations. Finney, C., Curtis, J. R., Paulus, H. E., Mudano, A., Pisu, M.,
Arthritis & Rheumatism, 54, 226–229. Elkins-Melton, M., Outman, R., Allison, J. J., Suarez
Jacobsen, E. (1932). Electrophysiology of mental activ- Almazor, M., Bridges S. L. Jr., Chatham, W. W., Hochberg, M.,
ities. American Journal of Psychology, 44, 677–694. MacLean, C., Mikuls, T., Moreland, L. W., O’Dell, J.,
Jensen, M. P. (2011). Psychosocial approaches to pain man- Turkiewicz, A. M., Furst, D. E., & American College of
agement: An organizational framework. Pain, 152(4), 717–725. Rheumatology (2008). American College of Rheumatology
Kaplan, M. S., Huguet, N., Newsom, J. T., & 2008 recommendations for the use of nonbiologic and
McFarland, B. H. (2003). Characteristics of physically inac- biologic disease-modifying antirheumatic drugs in
tive older adults with arthritis: results of a population-based rheumatoid arthritis. Arthritis Care & Research, 59(6),
study. Preventive Medicine, 37(1), 61–67. 762–784.
Lawrence, R. C., Felson, D. T., Helmick, C. G., Sacks, H., Chalmers, T. C., & Smith, H. (1982). Randomized
Arnold, L. M., Choi, H., Deyo, R. A., Gabriel, S., Hirsch, R., versus historical controls for clinical trials. American Jour-
Hochberg, M. C., Hunder, G. G., Jordan, J. M., Katz, J. N., nal of Medicine, 72(2), 233–240.
Kremers, H. M., Wolfe, F., & National Arthritis Data Work- Schoenfeld-Smith, K., Petroski, G. F., Hewett, J. E.,
group (2008). Estimates of the prevalence of arthritis and Johnson, J. C., Wright, G. E., Smarr, K. L., Walker, S. E., &
other rheumatic conditions in the United States. Part II. Parker, J. C. (1996). A biopsychosocial model of disability
Arthritis Rheumatism, 58(1), 26–35. in rheumatoid arthritis. Arthritis Care & Research, 9,
Lewandowski, W. A. (2004). Patterning of pain and power 368–375.
with guided imagery. Nursing Science Quarterly, 17(3), Schulz, K. F., Chalmers, L., Hayes, R. J., & Altman, D. G.
233–241. (1995). Empirical evidence of bias: Dimensions of method-
Lewandowski, W., Good, M., & Draucker, C. B. (2005). ological quality associated with estimates of treatment ef-
Changes in the meaning of pain with the use of guided im- fects in controlled trials. Journal of the American Medical
agery. Pain Management Nursing, 6(2), 58–67. Association, 273(5), 408–412.
Guided Imagery as Treatment for AORD 803

Shih, M., Hootman, J. M., Kruger, J., & Helmick, C. G. Retrieved from http://plato.stanford.edu/archives/fall2014/
(2006). Physical activity in men and women with arthritis: entries/mental-imagery/.
National Health Interview Survey, 2002. American Journal van Driel, M. L., de Sutter, A., Maeseneer, J. D., &
of Preventive Medicine, 30(5), 385–393. Christiaens, T. (2009). Searching for unpublished trials in
Taylor, S. D., Everett, S. V., Taylor, T. N., Watson, D. J., Cochrane reviews may not be worth the effort. Journal of
& Taylor-Stokes, G. (2013). A measure of treatment Clinical Epidemiology, 62(8), 838–844.
response: patient and physician satisfaction with Woolf, A. D., Zeidler, H., Haglund, U., Carr, A. J.,
traditional NSAIDs for osteoarthritis control. Open Chaussade, S., Cucinotta, D., Veale, D. J., & Martin-Mola, E.
Access Rheumatology: Research and Reviews, 5, (2004). Musculoskeletal pain in Europe: its impact and a
69–76. comparison of population and medical perceptions of
Thomas, N. J. T. (2014). Mental Imagery. In E. N. Zalta (Ed.), treatment in eight European countries. Annals of Rheu-
The Stanford encyclopedia of philosophy (Fall 2014 ed.). matic Diseases, 63, 342–347.