Patient Education and Counseling 98 (2015) 269–282
Contents lists available at ScienceDirect
Patient Education and Counseling journal homepage: www.elsevier.com/locate/pateducou
Review
Educational interventions for cancer pain. A systematic review of systematic reviews with nested narrative review of randomized controlled trials Rosalind Adam *, Christine Bond, Peter Murchie Centre of Academic Primary Care, University of Aberdeen, Aberdeen, UK
A R T I C L E I N F O
A B S T R A C T
Article history: Received 21 May 2014 Received in revised form 8 October 2014 Accepted 8 November 2014
Objectives: Educational interventions are one approach to improving cancer pain management. This review aims to determine whether educational interventions can improve cancer pain management and to characterize components of cancer pain educational interventions. Methods: Medline, EMBASE, CINAHL, and Cochrane databases were searched. Systematic reviews that assessed educational interventions to improve cancer pain management were included. Randomized controlled trials (RCTs) were identified from each review. A narrative approach was taken to summarizing the nature and components of interventions. Results: Eight systematic reviews and 34 randomized controlled trials (RCTs) were reviewed. Interventions targeting patients can achieve small to moderate reductions in pain intensity. Interventions targeting professionals can improve their knowledge but most trials have not assessed for resultant patient benefits. All interventions included at least one of seven core components: improving knowledge about the nature of cancer pain; aiding communication about cancer pain; enhancing pain assessment; improving analgesic prescribing; tackling barriers to analgesic nonadherence; teaching non-pharmacological pain management strategies; and promoting re-assessment. Conclusions: Cancer pain educational interventions can improve pain outcomes. They are complex heterogeneous interventions which often contain a combination of active components. Practice implications: Suggestions are made to aid the development of future interventions. ß 2014 Elsevier Ireland Ltd. All rights reserved.
Keywords: Cancer Pain Pain education Patient education Professional education Carer education
Contents 1. 2.
3.
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Inclusion and exclusion criteria . . . . . . . . . . . . . . . . . . . 2.1. Search strategy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2. Study selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3. Quality appraisal . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.4. Data extraction and synthesis . . . . . . . . . . . . . . . . . . . . 2.5. Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Characteristics of the reviews and their included RCTs 3.1. Quality of reviews. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2. Results of systematic reviews . . . . . . . . . . . . . . . . . . . . 3.3. Narrative description of individual interventions . . . . . 3.4.
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
* Corresponding author at: Centre of Academic Primary Care, University of Aberdeen, Room 1:130, Polwarth Building, Foresterhill, Aberdeen AB25 2ZD, UK. Tel.: +44 0 1224437906; fax: +44 0 1224437971. E-mail addresses: [email protected], [email protected] (R. Adam). http://dx.doi.org/10.1016/j.pec.2014.11.003 0738-3991/ß 2014 Elsevier Ireland Ltd. All rights reserved.
. . . . . . . . . . . .
. . . . . . . . . . . .
270 270 270 270 271 271 271 271 272 272 272 279
270
4. 5.
R. Adam et al. / Patient Education and Counseling 98 (2015) 269–282
Discussion and conclusions . Discussion . . . . . . . . . 4.1. Conclusion . . . . . . . . . . . . . . References . . . . . . . . . . . . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
1. Introduction Cancer incidence and prevalence rates are increasing in developed countries due to an aging population and improvements in screening and treatment [1]. Pain of moderate to severe intensity affects around 40% of cancer patients at diagnosis, rising to 70% or more at the end of life [2,3]. Cancer pain is burdensome at an individual and health care level [4]. For example, in Scotland, over 30% of primary care out of hours contacts for cancer symptoms relate to pain and around 14% of these patients are managed with education and advice [5]. Several large studies of emergency department visits by oncology patients in the USA and Canada have shown that pain is the most frequent reason for attendance, [6,7] and that around 30% could be avoided [8]. These figures suggest considerable and potentially preventable suffering and healthcare utilization due to poor cancer pain management despite effective treatments being available [9]. Suboptimal cancer pain management is likely to be a product of both patient and health care provider (HCP) behaviors. For example there is evidence that patients do not always report their pain to HCPs, [10] can have difficulty communicating with HCPs about cancer pain and associated symptoms [11], and intentionally or unintentionally fail to adhere to prescribed analgesic regimes [12]. Studies have identified various determinants of these behaviors: patients are less likely to report their pain to a HCP if they believe that cancer pain is inevitable and uncontrollable [13] and they may under-report pain to keep the clinician focused on other aspects of their condition during the clinical assessment [14]. Pain monitoring by patients can be problematic and subject to recall bias and underestimation, for example current pain ratings taken at an appointment do not correlate well with self-reported measures in daily or hourly pain diaries [15]. Patients’ beliefs and attitudes about analgesic side effects, addiction, and tolerance are known to influence medication adherence and subsequent pain control [16]. Intentional analgesic non-adherence can also occur if patients believe that pain is a helpful and necessary symptom, for example as an indicator of disease progression [17]. HCP behaviors may also prevent effective cancer pain management. It is known that analgesic prescriptions for patients with cancer pain are frequently suboptimal. Common deficiencies in analgesic prescriptions include failure to prescribe ‘‘around the clock’’ long acting analgesics, insufficient dosing of breakthrough medication, and failure to utilize adjuvant agents [18,19]. Determinants of these behaviors include inadequate knowledge about pain assessment and opioid prescribing [20,21] and concerns about hastening death with strong opioids [20]. Patient and HCP knowledge and attitudes appear to combine to influence patient-provider interactions and pain management behaviors. Interventions which target these behavioral determinants may therefore improve pain management. Interventions which have tried to improve health care provider or patient/carer knowledge about cancer pain management and improve pain management behaviors have broadly been referred to in the literature as educational interventions. Guidelines on the anticipatory management of cancer pain stress the importance of cancer pain education [22], however there is no recognized gold standard for cancer pain education in clinical practice and, as explained previously, cancer pain management is often suboptimal in clinical practice [9]. In the last 30 years a substantial number of
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
279 279 280 280
educational interventions aiming to improve the management of cancer pain have been subjected to clinical trials. This review aimed to synthesize this evidence and to explore whether educational interventions can improve pain outcomes. In addition this review aimed to describe the components of these interventions and the theoretical rationale for their mechanisms of action according to the original study authors. By summarizing what has been learned about cancer pain education and by identifying the key components of previous interventions, this review will inform the development of future educational interventions to tackle suboptimal cancer pain management behaviors. 2. Methods A systematic review of systematic reviews was conducted with nested review of the randomized controlled trials (RCTs) contained within selected reviews. Systematic reviews of systematic reviews are known to be useful, particularly for aiding policy and clinical decisions in healthcare when there are multiple existing reviews of a healthcare intervention [23]. By systematically identifying all reviews of cancer pain education, quality appraising existing reviews, and comparing and contrasting conclusions, it is possible to summarize the evidence base on educational interventions for cancer pain. This strategy was used to answer the research question: are educational interventions effective for improving pain outcomes in patients with cancer? Comprehensive evaluation of the reviews included a narrative summary of the randomized controlled trials underpinning them, the structure and components of each individual intervention and the rationale given by the authors for expected mechanism of action. 2.1. Inclusion and exclusion criteria This review considered systematic reviews that evaluated educational interventions for cancer pain management. Inclusion and exclusion criteria were formulated according to the ‘‘PICOS’’ method described by PRISMA [24]. Participants were adult patients, relatives, carers, or health professionals. Interventions were educational interventions in which participants received instruction, information about, or training in cancer pain reporting, assessment, and/or management. All publication dates were considered and publications in any language were considered, however, the search was conducted in English only. Reviews which included any experimental design were included. Reviews which included trials of cognitive behavioral therapy, mindfulness, relaxation, acupuncture, and hypnosis were excluded. Reviews which did not consider pain as an outcome measure were excluded. Reviews without a systematic methodological approach and editorials were excluded. Unpublished work and academic theses were not reviewed. 2.2. Search strategy Keywords and synonyms relating to cancer, education, and pain were agreed by RA and PM. Resulting keywords and Boolean operators were combined with the advice of an experienced medical librarian to search Medline, EMBASE, CINAHL, and Cochrane medical databases. The search strategy for Medline is presented in appendix 1. Computerized database searches were
R. Adam et al. / Patient Education and Counseling 98 (2015) 269–282
initially performed in October 2012 and a full Medline search was updated on 18th August 2013. Reference lists from all included reviews and RCTs were hand searched by RA. PROSPERO International prospective register of systematic reviews, Cochrane library, and clinicaltrials.gov were searched for in process citations of registered reviews. 2.3. Study selection Titles were screened independently by RA and CB against the inclusion criteria. Abstracts were retrieved of included titles, duplicate abstracts were removed, and remaining abstracts were assessed independently by RA and CB. Those judged by either author to meet the inclusion criteria were accessed in full text. Full text articles were reviewed independently by RA and CB, and inclusion was decided by discussion. 2.4. Quality appraisal Each review was assessed against the ‘‘Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) checklist by RA.
271
methods of assessing for bias, heterogeneity, and trial quality, the results of any meta-analysis; and conclusions. Full text was obtained for every RCT contained within the selected systematic reviews. A second data extraction pro-forma was created to collect data on: first author; date of publication; country of origin; the number of participants before randomization, nature of the intervention; nature of the comparator; what the authors considered to be the main theoretical or active components of the interventions; and outcomes. Educational interventions for cancer pain are complex interventions and analysis must take account of interactions between multiple components [25,26]. In addition to aggregating and summarizing the results of systematic reviews, we took an interpretative approach and produced a narrative synthesis of the RCTs contained within these reviews. RCTs were chosen because they were deemed to be the best quality evidence from the systematic reviews. The arguments and rationales provided by the authors for the mechanisms of their interventions were considered in order to elucidate the key components of these interventions and to summarize common and shared components of cancer pain educational interventions in general. 3. Results
2.5. Data extraction and synthesis
Idenficaon
A data extraction form was compiled using Microsoft Excel 2010 and used to record extracted data from reviews including: year of publication; journal of publication; aims; inclusion and exclusion criteria; total number of participants; primary and secondary outcome measures; nature of any statistical analyses;
A PRISMA diagram is shown in Fig. 1. 2066 titles were identified by searching four databases, hand-searching bibliographies of reviews and randomized controlled trials, and checking trial registries. 27 full text articles were assessed, of which eight reviews satisfied the inclusion and exclusion criteria and were included in the narrative synthesis [27–34].
Records identified through EMBASE CINAHL, MEDLINE, and Cochrane (n = 2054 )
Additional records identified through other sources (n =12 )
Screening
Titles screened (n = 2066 )
Abstracts identified (n = 74 )
Included
Eligibility
Abstracts screened (n=64)
Full-text articles assessed for eligibility (n = 27 )
Titles excluded based upon relevance (n=1992)
Duplicate abstracts removed (n=10)
Abstracts exluded, mainly because the study was not a systematic review or did not consider pain. (n=37)
Full-text articles excluded, after inclusion and exclusion criteria applied (n =19)
Studies included in narrative synthesis (n = 8 )
Fig. 1. PRISMA chart detailing study identification and selection process.
272
R. Adam et al. / Patient Education and Counseling 98 (2015) 269–282
3.1. Characteristics of the reviews and their included RCTs The main characteristics of the eight included reviews are summarized in Table 1. The reviews were published between 2001 and 2013 and were undertaken in Canada [27,31]; USA [28]; China [33]; Netherlands [30]; South Korea [44] and the United Kingdom [29]. Koller et al.’s review included collaboration between authors from Switzerland, Germany, and USA [32]. The reviews included studies published between 1982 and 2011. Three reviews [30,33,34] included only randomized controlled trials (RCTs) whilst the remaining five reviews also included other experimental study designs. All eight reviews included interventions targeting adult patients with cancer, two reviews also included studies of HCP education [27,28], and one review included studies of patient, HCP, and caregiver education [31]. A variety of pain outcomes were assessed including pain intensity, pain duration, average pain, worst pain, and pain interference. Quality of life was considered in four reviews [27,29,31,33]. The eight selected systematic reviews included 35 reports of 34 RCTs that had sample sizes before randomization ranging from 27 to 1256 participants with an average sample size of 208 participants [35–69]. Two trials were pilot studies [36,48]. The countries of origin were USA (24/34 trials), Taiwan (3/34 trials); Netherlands (2/34 trials), Australia (2/34 trials), Greece (1/34 trials), Canada (1/24 trials) and Turkey (1/24 trials). Twenty studies targeted patients alone, six studies targeted HCPs alone as participants, six studies targeted patients and their caregiver, one study targeted only carers, and one study included patients and HCPs. 32 of 34 interventions were for mixed cancer types. With respect to patient educational interventions, four interventions were exclusively applied to hospital inpatients [36,45,48,68]. The remainder targeted outpatients or ambulatory patients within the community. Interventions were mainly delivered by nurses. The most frequently used pain assessment methods were the visual analogue scale (VAS) and the Brief Pain Inventory (BPI). 3.2. Quality of reviews PRISMA 27 point quality checklists for each review are presented in Table 2. Meta-analyses by Bennett et al. [29] and Cummings et al. [31] scored highest for quality fulfilling 25 and 24 of the quality criteria, respectively. There were potential problems with study identification in four reviews by Goldberg et al. [28], Oldenmenger et al. [30], Ling et al. [33], and Jho et al. [34] either because the search strategies were not fully described or because restricted keywords were used. Reviews by Allard et al. [27], and Oldenmenger et al. [30] did not make it clear how studies were selected for inclusion, nor was it clear which authors or how many authors were involved in the selection process. Several reviews restricted their selection to studies reported in English language [27,28,30,33]. The earliest systematic reviews of cancer pain educational interventions [27,28] did not include any quality assessment of the included trials. All reviews published after 2007 included an assessment of trial quality and risk of bias. Reviews which reported on randomized controlled trials all described deficiencies in the reporting of power analysis, blinding, methods of randomization, and methods of dealing with drop-outs. 3.3. Results of systematic reviews Findings from each included review are summarized in Table 3. Allard et al. [27] identified 33 studies. Twenty five were studies of
educational interventions targeting HCPs, of which four were randomized controlled trials. Of 25 studies, only one made an assessment of the impact of professional education on the patient’s pain management. There was a consensus that educational interventions improved attitude and knowledge scores among health professionals but it was not clear whether this translated into a clinical benefit to patients. Eight studies targeted patients or family caregivers for pain education, and two of these were RCTs. Patient’s pain relief improved in five of these studies (including one RCT). Goldberg et al. [28] found no consistent improvement in patient’s pain management in studies of inpatient patient and HCP educational interventions. This was despite improvements in pain documentation, nurses’ knowledge and attitudes, and increased patient satisfaction with care. Oldenmenger et al. [30] identified 11 RCTs assessing patient education and all studies used pain intensity as the main outcome measure. Five of the eleven studies reported a statistically significant reduction in patient’s pain scores in the intervention groups. A statistically significant improvement was reported in eight out of ten studies for patient’s knowledge about cancer pain and for three of six studies for self-reported adherence to analgesics. Ling et al. [33] identified four RCTs of patient educational interventions. There was a statistically significant reduction in the usual severity of pain in one study, and the average intensity of pain in another study. Two of the four RCTs did not show any statistically significant difference in pain measures. None of the included studies demonstrated improvement in pain interference (how much pain interferes with daily activities) or quality of life measures. Three meta-analyses were conducted. Bennett et al. [29] included 15 studies and found a statistically significant benefit of education on average pain (weighted mean difference WMD 1.1, 95% confidence interval 1.8 to 0.41), maximum pain (WMD 0.78, 1.21 to 0.35), least pain (WMD 0.98, 1.68 to 0.28), and current pain intensity (WMD 0.65, 1.21 to 0.09). Overall there was an average of 1 point reduction on the 10 point Brief Pain Inventory (BPI) scale. Patient’s pain knowledge and attitudes were also improved by half a point on a five point scale (WMD 0.52, 95% CI 0.04 to 1.0). There were no significant improvements in mood, quality of life, or pain interference. Cummings et al. [31] included 11 studies of patient or family educational interventions including between two and six studies for each individual pain outcome. Results were reported as standardized mean differences (SMD), with positive values indicating a benefit of the intervention. There were two statistically significant results favoring pain education with respect to average pain intensity (SMD 0.43, 95% CI 0.13 to 0.74) and least pain (0.93, 95% CI 0.44 to 1.42). There was substantial heterogeneity affecting both these results. There were no significant effects of education on worst pain, overall pain, current pain, or pain interference. Jho et al. [34] included 12 studies in the meta-analysis and also reported SMD but with negative values indicating a beneficial effect of the intervention. Pain intensity was reduced in the intervention groups compared to control (SMD 0.11 95%CI 0.20 to 0.02). There was no consensus about factors which might modify the effectiveness of educational interventions. Koller et al. [32] systematically identified potential content and structure components from 24 studies of patient educational interventions. The authors found that nine of 24 studies had statistically and clinically meaningful reductions in pain, but that no specific content or structure component was associated with success of the intervention.
Table 1 Characteristics of systematic reviews. Metaanalysis?
Dates of studies
Included study designs
Participants
Settings of included studies
Allard, 2001 [27]
No
1982 to 1999
All study designs except simple narrative descriptions of pain control interventions
Patients, carers, professionals
Inpatient and outpatient
Goldberg, 2007 [28]
No
1982 to 2006
Patients and professionals
Bennett, 2009 [29]
Yes
1987 to 2008
Meta-analyses, RCTs, controlled trials, pre-post test designs RCTs and controlled trials
Oldenmenger, 2009 [30]
No
1987 to 2005
Cummings, 2011 [31]
Yes
Ling, 2011 [33]
Studies
Sample size
Quality criteria
Primary outcome measures
33
849
None
Inpatient
21
None
Patients
Inpatient and outpatient
21
Included institutional interventions 3501
RCTs*
Patients and professionals
Inpatient and outpatient
16*
2182
Cochrane criteria
1987 to 2009
RCTs, controlled trials, interrupted time series studies, controlled before and after studies
Patients, carers, and professionals
Inpatient and outpatient
26
4893
Cochrane criteria
No
2004 to 2010
RCTs
Patients
Outpatient
647
Jadad system
Koller, 2012 [32]
No
1987 to 2010
RCTs and controlled trials
Patients
Inpatient and outpatient
24
4193
Cochrane criteria
Jho, 2013 [34]
Yes
1986 to 2011
RCTs
Patients
Inpatient and outpatient
12
2169
Jadad system
Attitudes and knowledge, pain management (including pain documentation, change in analgesic prescription, adherence, and pain management index), pain relief, quality of life Patient satisfaction, documentation of pain intensity, nurses knowledge and attitudes about pain, pain severity Pain intensity; knowledge and attitude; medication adherence and selfefficacy; mood; and quality of life Pain intensity, patient and professional knowledge, barriers, adherence to analgesics, pain management index Change in behavior (professional); pain interference; usual pain; worst pain; overall pain; current/present pain; quality of life, satisfaction with treatment, and change in pain treatment Quality of life, pain intensity, pain interference Qualitative content analysis to categorize the structure and content components of the interventions Pain intensity
4
Cochrane criteria
R. Adam et al. / Patient Education and Counseling 98 (2015) 269–282
First author, year published
*
Oldenmenger et al. conducted the review in two parts. The first part was a systematic literature review on barriers to cancer pain management and all study designs were included. The second part reviewed RCTs of interventions to overcome barriers to cancer pain management, namely patient and professional education, and interventions to improve pain assessment by professionals. For the purpose of this review, only the review of interventions is considered here.
273
R. Adam et al. / Patient Education and Counseling 98 (2015) 269–282
274
Table 2 PRISMA quality assessment check-lists for systematic reviews. Quality measure
Allard [27]
Goldberg [28]
Bennett [29]
Oldenmenger [30]
Cummings [31]
Ling [33]
Koller [32]
Jho [34]
Title identifies report as systematic review or meta-analysis Report contains structured summary Introduction contains rationale for review in context of what is already known Contains explicit statement of questions being addressed with reference to PICOS A protocol exists and the review has been registered All information sources are listed along with the date last searched Full electronic search strategy is presented for at least one database Process for selecting studies is listed Method of data extraction and process for confirming data from investigators is described Lists and defines all variables for which data sought including any assumptions Describes methods for assessing risk of bias in individual studies and how this is used in data synthesis States principal summary measures Describes methods of handling data and combining results of studies including measures of consistency Specifies any risk of bias that may affect cumulative evidence Describes any methods of additional analysis e.g. subgroup analysis and whether these were pre-specified Reports numbers of studies screened, assessed for eligibility, and included with reasons for exclusions at each stage Presents characteristics for which data were extracted for each individual study Presents risk of bias for each study Presents simple summary data for each intervention group with effect estimates and confidence intervals Presents results of each meta-analysis done, including confidence intervals and measures of consistency Presents results of any assessment of risk of bias across studies Gives results of additional analyses, if done Summarizes main findings and strength of findings and relevance for key groups Discusses limitations at study, outcome, and review level Provides general interpretation of the results in context and implications Describes source of funding
H H H
H H H
H H H
H x H
x H H
H H H
H H H
H H H
x
x
H
x
H
H
H
H
x H H x x
x H x x U
x H U U U
x H x x x
H H U U U
x H x U U
x H U U U
x H x U U
U
x
U
x
U
U
U
U
U
x
U
U
U
U
U
U
n/a n/a
n/a n/a
U U
n/a n/a
U U
n/a U
U U
U U
U n/a
x n/a
U U
x n/a
U n/a
n/a n/a
x U
x U
U
x
U
x
U
U
x
U
U
U
U
U
U
U
U
U
x x
n/a
U U
U n/a
U U
U n/a
n/a
U U
n/a
n/a
U
n/a
U
n/a
n/a
U
n/a n/a U
x n/a U
U U U
n/a n/a U
U U U
n/a n/a U
n/a U U
x U U
U U
x U
U U
x U
U U
U U
U U
U U
U
U
U
x
U
U
U
x
Allard et al. [27] and Bennett et al. [29] found that a single educational exposure was not inferior to multiple episodes of education, whilst Cummings et al. [31] suggested that high dose, high intensity interventions yielded proportionately more positive results. Jho et al. [34] found that baseline patient characteristics
were important in determining the success of educational interventions and that the benefit of education seemed to diminish over time. Beneficial effects of education were more pronounced in those with a prognosis of more than three months, inpatients, and those
Table 3 Results of systematic reviews. First author, year
Country of review
Study dates
Summary of findings
Allard, 2001[27]
Canada
1982–1999
Goldberg, 2007 [28]
USA
1982–2006
Bennett, 2009 [29]
UK
1987–2008
Oldenmenger, 2009 [30]
Netherlands
1987–2005
Cummings, 2011[31]
Canada
1987–2009
Ling, 2011 [33]
Hong Kong
2004–2010
Koller, 2012 [32]
Switzerland, USA, Germany
1987–2010
Jho, 2013 [34]
South Korea
1986–2011
Interventions targeting health professionals can improve their knowledge and attitudes toward cancer pain. It is unclear whether this translates into improvements in patient’s pain levels. Patient interventions provided by nurses may be effective in improving patient’s attitudes and knowledge, and using a daily pain diary to aid communication about pain intensity can lead to reduced patient reported pain scores Looked only at inpatient setting Professional and patient education led to increased patient satisfaction, increased documentation of pain intensity, improvements in nurses’ knowledge and attitudes but did not consistently improve pain intensity Patient based educational interventions improved patient’s knowledge and attitudes and reduced pain intensity, but did not reduce interference from pain Equivocal results. Patient outcomes were not assessed in studies of professional education. Patient education reduced patient pain intensity in seven of eleven RCTs Knowledge translation interventions targeting health care providers, patients, and carers often, but not always, improved knowledge, skills, and attitudes about pain control and pain outcomes Educational interventions targeting patients did not improve quality of life. RCT results on pain intensity were equivocal with two positive trial results and two negative results Nine of 24 studies of educational interventions targeting patients had statistically and clinically meaningful reductions in pain. No specific structure or content component was particularly associated with statistically or clinically meaningful results Patient educational interventions resulted in a small overall reduction in pain intensity
Table 4 Summary of randomized controlled trials contained within included reviews. First author
Comparator
Anderson [35]
Culture specific video and booklet on pain management for African American and Hispanic patients with baseline cancer pain
Video on nutrition
x
x
Chang [36]
16 page pain educational booklet addressing recognized barriers to successful pain management, discussed face to face with in patients with cancer related pain
Usual care
x
x
Clotfelter [37]
14 min video and booklet addressing cancer pain management principles, delivered to elderly patients (>65 years) with cancer 60 years and carers, written booklet, audiocassette, written instructions for nonpharmacological management Three 40–60 min face to face educational sessions for patients and their partners in their home over two weeks, supplemented with written information, audiotapes, and a videotape
Intervention components Pain information
Usual care
x
Usual care
x
Communication
Outcomes Assessment
Prescribing
Analgesic adherence
Nonpharmacological
Reassessment
Statistically significant reduction in worst pain in education group compared to control group two to four weeks after the intervention but this was not maintained. There was no difference between groups for quality of life or functional measurements Intervention group had statistically significant reductions in barriers questionnaire scores and self -reported medication adherence compared to control, and statistically significant reductions in pre- intervention/postintervention pain intensity and pain interference measurements Statistically significant reduction in pain intensity in the intervention group compared to the control group at two weeks follow up
x
x
x
x
x
Usual care
x
x
Usual care in control communities
x
x
Usual care
x
x
Usual care
x
x
x
x
x
x
x
x
x
x
Statistically significant improvement in patient knowledge in the educational intervention group compared to control, but no statistically significant changes in pain measures Improvement in health professional knowledge after intervention. Intervention did not lead to statistically significant improvement in patient pain scores
Patients in the pain management algorithm group had statistically significant reductions in usual pain compared to control. There was no significant difference in quality of life between the groups No statistically significant improvements in patient pain or knowledge scores between intervention and control groups Patients receiving the educational intervention had statistically significant reductions in pain intensity and increases in self- reported medication adherence at four weeks follow up Patients in the intervention group had no significant improvement in pain or quality of life measurements
R. Adam et al. / Patient Education and Counseling 98 (2015) 269–282
Nature of intervention
275
276
Table 4 (Continued ) First author
Kravitz [45]
Kravitz [46]
Kroenke [47]
Lin [49]
Lovell [50]
McDonald [51]
McMillan [52]
Comparator
Use of a pain assessment sheet to display pain scores on a hospital ward Face to face tailored education and coaching about pain misconceptions and communicating about pain with physicians
Periodic pain assessment, (not displayed) Usual care and a booklet about cancer pain
Five telephone calls from a specialist nurse who assessed symptoms, adherence, and educated about pain and depression Medications altered according to a guideline Automated patient reports of symptoms via internet or telephone could trigger additional input Hospital inpatients provided with an informational booklet with 15 min face to face education on the content daily for 5 days
Usual care
Provision of a 16 page educational booklet addressing common concerns about reporting pain and using analgesics delivered to outpatients with cancer pain and their family caregivers, content delivered face to face Provision of booklet and use of educational video. The content of the intervention was not clear from the reports (and attempts to contact the authors unsuccessful)
Email to community nurse requesting comprehensive pain assessment, provision of written pain educational materials to patients, and availability of a specialist oncology nurse to offer advice to community nurse Carers educated on ‘‘COPE’’ symptom management education—creativity, optimism, planning, expert intervention
Intervention components Pain information
Communication
Outcomes Assessment
Prescribing
x
Analgesic adherence
Nonpharmacological
Reassessment
x
No significant improvement in pain scores in the intervention group
x
x
x
x
Usual care
x
x
Usual care
x
x
x
x
x
x
Those in the intervention group had statistically significant reductions in average pain intensity, negative beliefs about opioids, and pain catastrophizing scores, and increased sense of control over pain compared to the control group Statistically significant reductions in barriers scores, pain interference score at 4 weeks, and self- reported adherence to medication regime in the intervention group compared to control
x
Usual care
Email reminder to nurse about assessing pain
Usual care, and usual care plus friendly visits
x
x
x
x
Tailored education and coaching improved communication scores, but not pain severity compared to the control group There were early improvements in pain interference, not sustained at 6 and 12 weeks Statistically significant reduction in pain severity and depression in the intervention group compared to usual care patients
x
Statistically significant reduction in average pain scores when booklet and video were used together. Marginal improvements in intervention groups when booklet and video were used separately (not statistically significant). No difference in quality of life or pain interference Statistically significant reduction in worst pain intensity in the augmented intervention group compared to the basic email reminder (control) group
No statistically significant reductions in pain or quality of life in the intervention group compared to control groups. One measure of patient distress was significantly reduced in the intervention group compared to controls
R. Adam et al. / Patient Education and Counseling 98 (2015) 269–282
Lai [48]
Nature of intervention
Miaskows ki [53]
Usual care
x
Standardized instruction on controlling pain
x
Van der Peet [61]
Use of de Wit’s pain management program [40]
Usual care
Ward [62]
Face to face tailored information about barriers to analgesic use provided to women with gynecological cancers, provision of a written booklet, and telephone call one week later to reinforce
Usual care
Oliver [54]
Patiraki [55]
Rimer [56]
Syrjala [57]
Trowbridge [58]
Vallerand [59]
Vallieres [60]
‘‘Sham’’ videos on Florence Nightingale (same duration)
x
x
x
x
No summary available for professionals
x
Statistically significant reductions in usual pain in intervention group compared to control. No statistically significant improvement in functional impairment, pain related knowledge, or pain frequency Videos improved professional’s knowledge test scores. No patient outcome assessment
x
Statistically significant improvements in self-reported analgesic adherence, no statistically significant reduction in pain intensity between the intervention and control groups
x
Statistically significant reductions in barriers scores, usual pain, and increases in self -reported opioid use in the intervention group compared to controls Patients with a pain summary sheet were more likely to have their analgesic regime changed. This did not translate into improvements in pain intensity Statistically significant increases in nurses’ knowledge scores, no assessment of patient outcome measures
x
x
x
x
Usual care
x
x
x
x
x
x
x
Pain intensity decreased significantly from baseline in the intervention group compared to the control group.
x
Explanation of the study but no intervention
x
x
x
Usual care
Information on nutrition
x
x
x
No significant difference in pain scores at 2 weeks, but significantly reduced average and worst pain levels at 3 weeks in intervention group compared to controls Knowledge about pain significantly increased in intervention group compared to controls. Pain intensity significantly reduced at four weeks in the intervention group, but not at eight weeks. No significant changes in quality of life between the groups Intervention and control groups both had improved pain measurements over time, but there were no significant differences in outcome measures between the intervention and control groups
R. Adam et al. / Patient Education and Counseling 98 (2015) 269–282
PROSELF intervention. Patient education on how to use a pillbox, how to communicate pain, and the need for regular changes in analgesia 20 min individualized educational and coaching session targeting knowledge about pain selfmanagement, communicating pain, and redressing personal misconceptions Four educational videos for nurses with the aim of improving their knowledge about nursing assessment of cancer pain, the main analgesic groups, opioid analgesics, and under-treated pain Printed educational booklet ‘‘no more pain’’, addressing myths about analgesics and rationale for adhering. Content delivered by nurse over 15 min. Nurse personalized aspects of the booklet to the patient’s pain control regime Educational video, provision of informational booklet. Content explained by nurse who reviewed individual concerns Patient reports on written pain assessment sheet with a summary for professionals highlighting poorly controlled pain Nurse education: four hour lectures with notes on analgesics, misconceptions, and communicating with physician colleagues Provision of information booklet, completion of pain diary, and increased access to a physician who could adjust analgesic
277
278
Table 4 (Continued ) First author
Ward [63]
Ward [64]
Wells [66]
Wilkie [67]
Yates [68]
Yildrim [69]
Comparator
Intervention components
Individually tailored intervention targeting patient attitudinal barriers toward reporting pain and using analgesics. Intervention delivered by telephone followed up with printed mailed copy of educational messages ‘‘RID’’ pain intervention, face to face 20 to 60 min education tackling pain management misconceptions
Usual care
x
x
No improvement in pain outcomes despite reduction in patient barriers questionnaire scores
Booklet about analgesics and misconceptions, follow up telephone call from nurse
x
x
RID pain intervention targeting individual patients or dyad of patient plus carer/significant other 15 min video for patients and carers,
Contents lists available at ScienceDirect
Patient Education and Counseling journal homepage: www.elsevier.com/locate/pateducou
Review
Educational interventions for cancer pain. A systematic review of systematic reviews with nested narrative review of randomized controlled trials Rosalind Adam *, Christine Bond, Peter Murchie Centre of Academic Primary Care, University of Aberdeen, Aberdeen, UK
A R T I C L E I N F O
A B S T R A C T
Article history: Received 21 May 2014 Received in revised form 8 October 2014 Accepted 8 November 2014
Objectives: Educational interventions are one approach to improving cancer pain management. This review aims to determine whether educational interventions can improve cancer pain management and to characterize components of cancer pain educational interventions. Methods: Medline, EMBASE, CINAHL, and Cochrane databases were searched. Systematic reviews that assessed educational interventions to improve cancer pain management were included. Randomized controlled trials (RCTs) were identified from each review. A narrative approach was taken to summarizing the nature and components of interventions. Results: Eight systematic reviews and 34 randomized controlled trials (RCTs) were reviewed. Interventions targeting patients can achieve small to moderate reductions in pain intensity. Interventions targeting professionals can improve their knowledge but most trials have not assessed for resultant patient benefits. All interventions included at least one of seven core components: improving knowledge about the nature of cancer pain; aiding communication about cancer pain; enhancing pain assessment; improving analgesic prescribing; tackling barriers to analgesic nonadherence; teaching non-pharmacological pain management strategies; and promoting re-assessment. Conclusions: Cancer pain educational interventions can improve pain outcomes. They are complex heterogeneous interventions which often contain a combination of active components. Practice implications: Suggestions are made to aid the development of future interventions. ß 2014 Elsevier Ireland Ltd. All rights reserved.
Keywords: Cancer Pain Pain education Patient education Professional education Carer education
Contents 1. 2.
3.
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Inclusion and exclusion criteria . . . . . . . . . . . . . . . . . . . 2.1. Search strategy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2. Study selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3. Quality appraisal . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.4. Data extraction and synthesis . . . . . . . . . . . . . . . . . . . . 2.5. Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Characteristics of the reviews and their included RCTs 3.1. Quality of reviews. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2. Results of systematic reviews . . . . . . . . . . . . . . . . . . . . 3.3. Narrative description of individual interventions . . . . . 3.4.
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
* Corresponding author at: Centre of Academic Primary Care, University of Aberdeen, Room 1:130, Polwarth Building, Foresterhill, Aberdeen AB25 2ZD, UK. Tel.: +44 0 1224437906; fax: +44 0 1224437971. E-mail addresses: [email protected], [email protected] (R. Adam). http://dx.doi.org/10.1016/j.pec.2014.11.003 0738-3991/ß 2014 Elsevier Ireland Ltd. All rights reserved.
. . . . . . . . . . . .
. . . . . . . . . . . .
270 270 270 270 271 271 271 271 272 272 272 279
270
4. 5.
R. Adam et al. / Patient Education and Counseling 98 (2015) 269–282
Discussion and conclusions . Discussion . . . . . . . . . 4.1. Conclusion . . . . . . . . . . . . . . References . . . . . . . . . . . . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
1. Introduction Cancer incidence and prevalence rates are increasing in developed countries due to an aging population and improvements in screening and treatment [1]. Pain of moderate to severe intensity affects around 40% of cancer patients at diagnosis, rising to 70% or more at the end of life [2,3]. Cancer pain is burdensome at an individual and health care level [4]. For example, in Scotland, over 30% of primary care out of hours contacts for cancer symptoms relate to pain and around 14% of these patients are managed with education and advice [5]. Several large studies of emergency department visits by oncology patients in the USA and Canada have shown that pain is the most frequent reason for attendance, [6,7] and that around 30% could be avoided [8]. These figures suggest considerable and potentially preventable suffering and healthcare utilization due to poor cancer pain management despite effective treatments being available [9]. Suboptimal cancer pain management is likely to be a product of both patient and health care provider (HCP) behaviors. For example there is evidence that patients do not always report their pain to HCPs, [10] can have difficulty communicating with HCPs about cancer pain and associated symptoms [11], and intentionally or unintentionally fail to adhere to prescribed analgesic regimes [12]. Studies have identified various determinants of these behaviors: patients are less likely to report their pain to a HCP if they believe that cancer pain is inevitable and uncontrollable [13] and they may under-report pain to keep the clinician focused on other aspects of their condition during the clinical assessment [14]. Pain monitoring by patients can be problematic and subject to recall bias and underestimation, for example current pain ratings taken at an appointment do not correlate well with self-reported measures in daily or hourly pain diaries [15]. Patients’ beliefs and attitudes about analgesic side effects, addiction, and tolerance are known to influence medication adherence and subsequent pain control [16]. Intentional analgesic non-adherence can also occur if patients believe that pain is a helpful and necessary symptom, for example as an indicator of disease progression [17]. HCP behaviors may also prevent effective cancer pain management. It is known that analgesic prescriptions for patients with cancer pain are frequently suboptimal. Common deficiencies in analgesic prescriptions include failure to prescribe ‘‘around the clock’’ long acting analgesics, insufficient dosing of breakthrough medication, and failure to utilize adjuvant agents [18,19]. Determinants of these behaviors include inadequate knowledge about pain assessment and opioid prescribing [20,21] and concerns about hastening death with strong opioids [20]. Patient and HCP knowledge and attitudes appear to combine to influence patient-provider interactions and pain management behaviors. Interventions which target these behavioral determinants may therefore improve pain management. Interventions which have tried to improve health care provider or patient/carer knowledge about cancer pain management and improve pain management behaviors have broadly been referred to in the literature as educational interventions. Guidelines on the anticipatory management of cancer pain stress the importance of cancer pain education [22], however there is no recognized gold standard for cancer pain education in clinical practice and, as explained previously, cancer pain management is often suboptimal in clinical practice [9]. In the last 30 years a substantial number of
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
279 279 280 280
educational interventions aiming to improve the management of cancer pain have been subjected to clinical trials. This review aimed to synthesize this evidence and to explore whether educational interventions can improve pain outcomes. In addition this review aimed to describe the components of these interventions and the theoretical rationale for their mechanisms of action according to the original study authors. By summarizing what has been learned about cancer pain education and by identifying the key components of previous interventions, this review will inform the development of future educational interventions to tackle suboptimal cancer pain management behaviors. 2. Methods A systematic review of systematic reviews was conducted with nested review of the randomized controlled trials (RCTs) contained within selected reviews. Systematic reviews of systematic reviews are known to be useful, particularly for aiding policy and clinical decisions in healthcare when there are multiple existing reviews of a healthcare intervention [23]. By systematically identifying all reviews of cancer pain education, quality appraising existing reviews, and comparing and contrasting conclusions, it is possible to summarize the evidence base on educational interventions for cancer pain. This strategy was used to answer the research question: are educational interventions effective for improving pain outcomes in patients with cancer? Comprehensive evaluation of the reviews included a narrative summary of the randomized controlled trials underpinning them, the structure and components of each individual intervention and the rationale given by the authors for expected mechanism of action. 2.1. Inclusion and exclusion criteria This review considered systematic reviews that evaluated educational interventions for cancer pain management. Inclusion and exclusion criteria were formulated according to the ‘‘PICOS’’ method described by PRISMA [24]. Participants were adult patients, relatives, carers, or health professionals. Interventions were educational interventions in which participants received instruction, information about, or training in cancer pain reporting, assessment, and/or management. All publication dates were considered and publications in any language were considered, however, the search was conducted in English only. Reviews which included any experimental design were included. Reviews which included trials of cognitive behavioral therapy, mindfulness, relaxation, acupuncture, and hypnosis were excluded. Reviews which did not consider pain as an outcome measure were excluded. Reviews without a systematic methodological approach and editorials were excluded. Unpublished work and academic theses were not reviewed. 2.2. Search strategy Keywords and synonyms relating to cancer, education, and pain were agreed by RA and PM. Resulting keywords and Boolean operators were combined with the advice of an experienced medical librarian to search Medline, EMBASE, CINAHL, and Cochrane medical databases. The search strategy for Medline is presented in appendix 1. Computerized database searches were
R. Adam et al. / Patient Education and Counseling 98 (2015) 269–282
initially performed in October 2012 and a full Medline search was updated on 18th August 2013. Reference lists from all included reviews and RCTs were hand searched by RA. PROSPERO International prospective register of systematic reviews, Cochrane library, and clinicaltrials.gov were searched for in process citations of registered reviews. 2.3. Study selection Titles were screened independently by RA and CB against the inclusion criteria. Abstracts were retrieved of included titles, duplicate abstracts were removed, and remaining abstracts were assessed independently by RA and CB. Those judged by either author to meet the inclusion criteria were accessed in full text. Full text articles were reviewed independently by RA and CB, and inclusion was decided by discussion. 2.4. Quality appraisal Each review was assessed against the ‘‘Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) checklist by RA.
271
methods of assessing for bias, heterogeneity, and trial quality, the results of any meta-analysis; and conclusions. Full text was obtained for every RCT contained within the selected systematic reviews. A second data extraction pro-forma was created to collect data on: first author; date of publication; country of origin; the number of participants before randomization, nature of the intervention; nature of the comparator; what the authors considered to be the main theoretical or active components of the interventions; and outcomes. Educational interventions for cancer pain are complex interventions and analysis must take account of interactions between multiple components [25,26]. In addition to aggregating and summarizing the results of systematic reviews, we took an interpretative approach and produced a narrative synthesis of the RCTs contained within these reviews. RCTs were chosen because they were deemed to be the best quality evidence from the systematic reviews. The arguments and rationales provided by the authors for the mechanisms of their interventions were considered in order to elucidate the key components of these interventions and to summarize common and shared components of cancer pain educational interventions in general. 3. Results
2.5. Data extraction and synthesis
Idenficaon
A data extraction form was compiled using Microsoft Excel 2010 and used to record extracted data from reviews including: year of publication; journal of publication; aims; inclusion and exclusion criteria; total number of participants; primary and secondary outcome measures; nature of any statistical analyses;
A PRISMA diagram is shown in Fig. 1. 2066 titles were identified by searching four databases, hand-searching bibliographies of reviews and randomized controlled trials, and checking trial registries. 27 full text articles were assessed, of which eight reviews satisfied the inclusion and exclusion criteria and were included in the narrative synthesis [27–34].
Records identified through EMBASE CINAHL, MEDLINE, and Cochrane (n = 2054 )
Additional records identified through other sources (n =12 )
Screening
Titles screened (n = 2066 )
Abstracts identified (n = 74 )
Included
Eligibility
Abstracts screened (n=64)
Full-text articles assessed for eligibility (n = 27 )
Titles excluded based upon relevance (n=1992)
Duplicate abstracts removed (n=10)
Abstracts exluded, mainly because the study was not a systematic review or did not consider pain. (n=37)
Full-text articles excluded, after inclusion and exclusion criteria applied (n =19)
Studies included in narrative synthesis (n = 8 )
Fig. 1. PRISMA chart detailing study identification and selection process.
272
R. Adam et al. / Patient Education and Counseling 98 (2015) 269–282
3.1. Characteristics of the reviews and their included RCTs The main characteristics of the eight included reviews are summarized in Table 1. The reviews were published between 2001 and 2013 and were undertaken in Canada [27,31]; USA [28]; China [33]; Netherlands [30]; South Korea [44] and the United Kingdom [29]. Koller et al.’s review included collaboration between authors from Switzerland, Germany, and USA [32]. The reviews included studies published between 1982 and 2011. Three reviews [30,33,34] included only randomized controlled trials (RCTs) whilst the remaining five reviews also included other experimental study designs. All eight reviews included interventions targeting adult patients with cancer, two reviews also included studies of HCP education [27,28], and one review included studies of patient, HCP, and caregiver education [31]. A variety of pain outcomes were assessed including pain intensity, pain duration, average pain, worst pain, and pain interference. Quality of life was considered in four reviews [27,29,31,33]. The eight selected systematic reviews included 35 reports of 34 RCTs that had sample sizes before randomization ranging from 27 to 1256 participants with an average sample size of 208 participants [35–69]. Two trials were pilot studies [36,48]. The countries of origin were USA (24/34 trials), Taiwan (3/34 trials); Netherlands (2/34 trials), Australia (2/34 trials), Greece (1/34 trials), Canada (1/24 trials) and Turkey (1/24 trials). Twenty studies targeted patients alone, six studies targeted HCPs alone as participants, six studies targeted patients and their caregiver, one study targeted only carers, and one study included patients and HCPs. 32 of 34 interventions were for mixed cancer types. With respect to patient educational interventions, four interventions were exclusively applied to hospital inpatients [36,45,48,68]. The remainder targeted outpatients or ambulatory patients within the community. Interventions were mainly delivered by nurses. The most frequently used pain assessment methods were the visual analogue scale (VAS) and the Brief Pain Inventory (BPI). 3.2. Quality of reviews PRISMA 27 point quality checklists for each review are presented in Table 2. Meta-analyses by Bennett et al. [29] and Cummings et al. [31] scored highest for quality fulfilling 25 and 24 of the quality criteria, respectively. There were potential problems with study identification in four reviews by Goldberg et al. [28], Oldenmenger et al. [30], Ling et al. [33], and Jho et al. [34] either because the search strategies were not fully described or because restricted keywords were used. Reviews by Allard et al. [27], and Oldenmenger et al. [30] did not make it clear how studies were selected for inclusion, nor was it clear which authors or how many authors were involved in the selection process. Several reviews restricted their selection to studies reported in English language [27,28,30,33]. The earliest systematic reviews of cancer pain educational interventions [27,28] did not include any quality assessment of the included trials. All reviews published after 2007 included an assessment of trial quality and risk of bias. Reviews which reported on randomized controlled trials all described deficiencies in the reporting of power analysis, blinding, methods of randomization, and methods of dealing with drop-outs. 3.3. Results of systematic reviews Findings from each included review are summarized in Table 3. Allard et al. [27] identified 33 studies. Twenty five were studies of
educational interventions targeting HCPs, of which four were randomized controlled trials. Of 25 studies, only one made an assessment of the impact of professional education on the patient’s pain management. There was a consensus that educational interventions improved attitude and knowledge scores among health professionals but it was not clear whether this translated into a clinical benefit to patients. Eight studies targeted patients or family caregivers for pain education, and two of these were RCTs. Patient’s pain relief improved in five of these studies (including one RCT). Goldberg et al. [28] found no consistent improvement in patient’s pain management in studies of inpatient patient and HCP educational interventions. This was despite improvements in pain documentation, nurses’ knowledge and attitudes, and increased patient satisfaction with care. Oldenmenger et al. [30] identified 11 RCTs assessing patient education and all studies used pain intensity as the main outcome measure. Five of the eleven studies reported a statistically significant reduction in patient’s pain scores in the intervention groups. A statistically significant improvement was reported in eight out of ten studies for patient’s knowledge about cancer pain and for three of six studies for self-reported adherence to analgesics. Ling et al. [33] identified four RCTs of patient educational interventions. There was a statistically significant reduction in the usual severity of pain in one study, and the average intensity of pain in another study. Two of the four RCTs did not show any statistically significant difference in pain measures. None of the included studies demonstrated improvement in pain interference (how much pain interferes with daily activities) or quality of life measures. Three meta-analyses were conducted. Bennett et al. [29] included 15 studies and found a statistically significant benefit of education on average pain (weighted mean difference WMD 1.1, 95% confidence interval 1.8 to 0.41), maximum pain (WMD 0.78, 1.21 to 0.35), least pain (WMD 0.98, 1.68 to 0.28), and current pain intensity (WMD 0.65, 1.21 to 0.09). Overall there was an average of 1 point reduction on the 10 point Brief Pain Inventory (BPI) scale. Patient’s pain knowledge and attitudes were also improved by half a point on a five point scale (WMD 0.52, 95% CI 0.04 to 1.0). There were no significant improvements in mood, quality of life, or pain interference. Cummings et al. [31] included 11 studies of patient or family educational interventions including between two and six studies for each individual pain outcome. Results were reported as standardized mean differences (SMD), with positive values indicating a benefit of the intervention. There were two statistically significant results favoring pain education with respect to average pain intensity (SMD 0.43, 95% CI 0.13 to 0.74) and least pain (0.93, 95% CI 0.44 to 1.42). There was substantial heterogeneity affecting both these results. There were no significant effects of education on worst pain, overall pain, current pain, or pain interference. Jho et al. [34] included 12 studies in the meta-analysis and also reported SMD but with negative values indicating a beneficial effect of the intervention. Pain intensity was reduced in the intervention groups compared to control (SMD 0.11 95%CI 0.20 to 0.02). There was no consensus about factors which might modify the effectiveness of educational interventions. Koller et al. [32] systematically identified potential content and structure components from 24 studies of patient educational interventions. The authors found that nine of 24 studies had statistically and clinically meaningful reductions in pain, but that no specific content or structure component was associated with success of the intervention.
Table 1 Characteristics of systematic reviews. Metaanalysis?
Dates of studies
Included study designs
Participants
Settings of included studies
Allard, 2001 [27]
No
1982 to 1999
All study designs except simple narrative descriptions of pain control interventions
Patients, carers, professionals
Inpatient and outpatient
Goldberg, 2007 [28]
No
1982 to 2006
Patients and professionals
Bennett, 2009 [29]
Yes
1987 to 2008
Meta-analyses, RCTs, controlled trials, pre-post test designs RCTs and controlled trials
Oldenmenger, 2009 [30]
No
1987 to 2005
Cummings, 2011 [31]
Yes
Ling, 2011 [33]
Studies
Sample size
Quality criteria
Primary outcome measures
33
849
None
Inpatient
21
None
Patients
Inpatient and outpatient
21
Included institutional interventions 3501
RCTs*
Patients and professionals
Inpatient and outpatient
16*
2182
Cochrane criteria
1987 to 2009
RCTs, controlled trials, interrupted time series studies, controlled before and after studies
Patients, carers, and professionals
Inpatient and outpatient
26
4893
Cochrane criteria
No
2004 to 2010
RCTs
Patients
Outpatient
647
Jadad system
Koller, 2012 [32]
No
1987 to 2010
RCTs and controlled trials
Patients
Inpatient and outpatient
24
4193
Cochrane criteria
Jho, 2013 [34]
Yes
1986 to 2011
RCTs
Patients
Inpatient and outpatient
12
2169
Jadad system
Attitudes and knowledge, pain management (including pain documentation, change in analgesic prescription, adherence, and pain management index), pain relief, quality of life Patient satisfaction, documentation of pain intensity, nurses knowledge and attitudes about pain, pain severity Pain intensity; knowledge and attitude; medication adherence and selfefficacy; mood; and quality of life Pain intensity, patient and professional knowledge, barriers, adherence to analgesics, pain management index Change in behavior (professional); pain interference; usual pain; worst pain; overall pain; current/present pain; quality of life, satisfaction with treatment, and change in pain treatment Quality of life, pain intensity, pain interference Qualitative content analysis to categorize the structure and content components of the interventions Pain intensity
4
Cochrane criteria
R. Adam et al. / Patient Education and Counseling 98 (2015) 269–282
First author, year published
*
Oldenmenger et al. conducted the review in two parts. The first part was a systematic literature review on barriers to cancer pain management and all study designs were included. The second part reviewed RCTs of interventions to overcome barriers to cancer pain management, namely patient and professional education, and interventions to improve pain assessment by professionals. For the purpose of this review, only the review of interventions is considered here.
273
R. Adam et al. / Patient Education and Counseling 98 (2015) 269–282
274
Table 2 PRISMA quality assessment check-lists for systematic reviews. Quality measure
Allard [27]
Goldberg [28]
Bennett [29]
Oldenmenger [30]
Cummings [31]
Ling [33]
Koller [32]
Jho [34]
Title identifies report as systematic review or meta-analysis Report contains structured summary Introduction contains rationale for review in context of what is already known Contains explicit statement of questions being addressed with reference to PICOS A protocol exists and the review has been registered All information sources are listed along with the date last searched Full electronic search strategy is presented for at least one database Process for selecting studies is listed Method of data extraction and process for confirming data from investigators is described Lists and defines all variables for which data sought including any assumptions Describes methods for assessing risk of bias in individual studies and how this is used in data synthesis States principal summary measures Describes methods of handling data and combining results of studies including measures of consistency Specifies any risk of bias that may affect cumulative evidence Describes any methods of additional analysis e.g. subgroup analysis and whether these were pre-specified Reports numbers of studies screened, assessed for eligibility, and included with reasons for exclusions at each stage Presents characteristics for which data were extracted for each individual study Presents risk of bias for each study Presents simple summary data for each intervention group with effect estimates and confidence intervals Presents results of each meta-analysis done, including confidence intervals and measures of consistency Presents results of any assessment of risk of bias across studies Gives results of additional analyses, if done Summarizes main findings and strength of findings and relevance for key groups Discusses limitations at study, outcome, and review level Provides general interpretation of the results in context and implications Describes source of funding
H H H
H H H
H H H
H x H
x H H
H H H
H H H
H H H
x
x
H
x
H
H
H
H
x H H x x
x H x x U
x H U U U
x H x x x
H H U U U
x H x U U
x H U U U
x H x U U
U
x
U
x
U
U
U
U
U
x
U
U
U
U
U
U
n/a n/a
n/a n/a
U U
n/a n/a
U U
n/a U
U U
U U
U n/a
x n/a
U U
x n/a
U n/a
n/a n/a
x U
x U
U
x
U
x
U
U
x
U
U
U
U
U
U
U
U
U
x x
n/a
U U
U n/a
U U
U n/a
n/a
U U
n/a
n/a
U
n/a
U
n/a
n/a
U
n/a n/a U
x n/a U
U U U
n/a n/a U
U U U
n/a n/a U
n/a U U
x U U
U U
x U
U U
x U
U U
U U
U U
U U
U
U
U
x
U
U
U
x
Allard et al. [27] and Bennett et al. [29] found that a single educational exposure was not inferior to multiple episodes of education, whilst Cummings et al. [31] suggested that high dose, high intensity interventions yielded proportionately more positive results. Jho et al. [34] found that baseline patient characteristics
were important in determining the success of educational interventions and that the benefit of education seemed to diminish over time. Beneficial effects of education were more pronounced in those with a prognosis of more than three months, inpatients, and those
Table 3 Results of systematic reviews. First author, year
Country of review
Study dates
Summary of findings
Allard, 2001[27]
Canada
1982–1999
Goldberg, 2007 [28]
USA
1982–2006
Bennett, 2009 [29]
UK
1987–2008
Oldenmenger, 2009 [30]
Netherlands
1987–2005
Cummings, 2011[31]
Canada
1987–2009
Ling, 2011 [33]
Hong Kong
2004–2010
Koller, 2012 [32]
Switzerland, USA, Germany
1987–2010
Jho, 2013 [34]
South Korea
1986–2011
Interventions targeting health professionals can improve their knowledge and attitudes toward cancer pain. It is unclear whether this translates into improvements in patient’s pain levels. Patient interventions provided by nurses may be effective in improving patient’s attitudes and knowledge, and using a daily pain diary to aid communication about pain intensity can lead to reduced patient reported pain scores Looked only at inpatient setting Professional and patient education led to increased patient satisfaction, increased documentation of pain intensity, improvements in nurses’ knowledge and attitudes but did not consistently improve pain intensity Patient based educational interventions improved patient’s knowledge and attitudes and reduced pain intensity, but did not reduce interference from pain Equivocal results. Patient outcomes were not assessed in studies of professional education. Patient education reduced patient pain intensity in seven of eleven RCTs Knowledge translation interventions targeting health care providers, patients, and carers often, but not always, improved knowledge, skills, and attitudes about pain control and pain outcomes Educational interventions targeting patients did not improve quality of life. RCT results on pain intensity were equivocal with two positive trial results and two negative results Nine of 24 studies of educational interventions targeting patients had statistically and clinically meaningful reductions in pain. No specific structure or content component was particularly associated with statistically or clinically meaningful results Patient educational interventions resulted in a small overall reduction in pain intensity
Table 4 Summary of randomized controlled trials contained within included reviews. First author
Comparator
Anderson [35]
Culture specific video and booklet on pain management for African American and Hispanic patients with baseline cancer pain
Video on nutrition
x
x
Chang [36]
16 page pain educational booklet addressing recognized barriers to successful pain management, discussed face to face with in patients with cancer related pain
Usual care
x
x
Clotfelter [37]
14 min video and booklet addressing cancer pain management principles, delivered to elderly patients (>65 years) with cancer 60 years and carers, written booklet, audiocassette, written instructions for nonpharmacological management Three 40–60 min face to face educational sessions for patients and their partners in their home over two weeks, supplemented with written information, audiotapes, and a videotape
Intervention components Pain information
Usual care
x
Usual care
x
Communication
Outcomes Assessment
Prescribing
Analgesic adherence
Nonpharmacological
Reassessment
Statistically significant reduction in worst pain in education group compared to control group two to four weeks after the intervention but this was not maintained. There was no difference between groups for quality of life or functional measurements Intervention group had statistically significant reductions in barriers questionnaire scores and self -reported medication adherence compared to control, and statistically significant reductions in pre- intervention/postintervention pain intensity and pain interference measurements Statistically significant reduction in pain intensity in the intervention group compared to the control group at two weeks follow up
x
x
x
x
x
Usual care
x
x
Usual care in control communities
x
x
Usual care
x
x
Usual care
x
x
x
x
x
x
x
x
x
x
Statistically significant improvement in patient knowledge in the educational intervention group compared to control, but no statistically significant changes in pain measures Improvement in health professional knowledge after intervention. Intervention did not lead to statistically significant improvement in patient pain scores
Patients in the pain management algorithm group had statistically significant reductions in usual pain compared to control. There was no significant difference in quality of life between the groups No statistically significant improvements in patient pain or knowledge scores between intervention and control groups Patients receiving the educational intervention had statistically significant reductions in pain intensity and increases in self- reported medication adherence at four weeks follow up Patients in the intervention group had no significant improvement in pain or quality of life measurements
R. Adam et al. / Patient Education and Counseling 98 (2015) 269–282
Nature of intervention
275
276
Table 4 (Continued ) First author
Kravitz [45]
Kravitz [46]
Kroenke [47]
Lin [49]
Lovell [50]
McDonald [51]
McMillan [52]
Comparator
Use of a pain assessment sheet to display pain scores on a hospital ward Face to face tailored education and coaching about pain misconceptions and communicating about pain with physicians
Periodic pain assessment, (not displayed) Usual care and a booklet about cancer pain
Five telephone calls from a specialist nurse who assessed symptoms, adherence, and educated about pain and depression Medications altered according to a guideline Automated patient reports of symptoms via internet or telephone could trigger additional input Hospital inpatients provided with an informational booklet with 15 min face to face education on the content daily for 5 days
Usual care
Provision of a 16 page educational booklet addressing common concerns about reporting pain and using analgesics delivered to outpatients with cancer pain and their family caregivers, content delivered face to face Provision of booklet and use of educational video. The content of the intervention was not clear from the reports (and attempts to contact the authors unsuccessful)
Email to community nurse requesting comprehensive pain assessment, provision of written pain educational materials to patients, and availability of a specialist oncology nurse to offer advice to community nurse Carers educated on ‘‘COPE’’ symptom management education—creativity, optimism, planning, expert intervention
Intervention components Pain information
Communication
Outcomes Assessment
Prescribing
x
Analgesic adherence
Nonpharmacological
Reassessment
x
No significant improvement in pain scores in the intervention group
x
x
x
x
Usual care
x
x
Usual care
x
x
x
x
x
x
Those in the intervention group had statistically significant reductions in average pain intensity, negative beliefs about opioids, and pain catastrophizing scores, and increased sense of control over pain compared to the control group Statistically significant reductions in barriers scores, pain interference score at 4 weeks, and self- reported adherence to medication regime in the intervention group compared to control
x
Usual care
Email reminder to nurse about assessing pain
Usual care, and usual care plus friendly visits
x
x
x
x
Tailored education and coaching improved communication scores, but not pain severity compared to the control group There were early improvements in pain interference, not sustained at 6 and 12 weeks Statistically significant reduction in pain severity and depression in the intervention group compared to usual care patients
x
Statistically significant reduction in average pain scores when booklet and video were used together. Marginal improvements in intervention groups when booklet and video were used separately (not statistically significant). No difference in quality of life or pain interference Statistically significant reduction in worst pain intensity in the augmented intervention group compared to the basic email reminder (control) group
No statistically significant reductions in pain or quality of life in the intervention group compared to control groups. One measure of patient distress was significantly reduced in the intervention group compared to controls
R. Adam et al. / Patient Education and Counseling 98 (2015) 269–282
Lai [48]
Nature of intervention
Miaskows ki [53]
Usual care
x
Standardized instruction on controlling pain
x
Van der Peet [61]
Use of de Wit’s pain management program [40]
Usual care
Ward [62]
Face to face tailored information about barriers to analgesic use provided to women with gynecological cancers, provision of a written booklet, and telephone call one week later to reinforce
Usual care
Oliver [54]
Patiraki [55]
Rimer [56]
Syrjala [57]
Trowbridge [58]
Vallerand [59]
Vallieres [60]
‘‘Sham’’ videos on Florence Nightingale (same duration)
x
x
x
x
No summary available for professionals
x
Statistically significant reductions in usual pain in intervention group compared to control. No statistically significant improvement in functional impairment, pain related knowledge, or pain frequency Videos improved professional’s knowledge test scores. No patient outcome assessment
x
Statistically significant improvements in self-reported analgesic adherence, no statistically significant reduction in pain intensity between the intervention and control groups
x
Statistically significant reductions in barriers scores, usual pain, and increases in self -reported opioid use in the intervention group compared to controls Patients with a pain summary sheet were more likely to have their analgesic regime changed. This did not translate into improvements in pain intensity Statistically significant increases in nurses’ knowledge scores, no assessment of patient outcome measures
x
x
x
x
Usual care
x
x
x
x
x
x
x
Pain intensity decreased significantly from baseline in the intervention group compared to the control group.
x
Explanation of the study but no intervention
x
x
x
Usual care
Information on nutrition
x
x
x
No significant difference in pain scores at 2 weeks, but significantly reduced average and worst pain levels at 3 weeks in intervention group compared to controls Knowledge about pain significantly increased in intervention group compared to controls. Pain intensity significantly reduced at four weeks in the intervention group, but not at eight weeks. No significant changes in quality of life between the groups Intervention and control groups both had improved pain measurements over time, but there were no significant differences in outcome measures between the intervention and control groups
R. Adam et al. / Patient Education and Counseling 98 (2015) 269–282
PROSELF intervention. Patient education on how to use a pillbox, how to communicate pain, and the need for regular changes in analgesia 20 min individualized educational and coaching session targeting knowledge about pain selfmanagement, communicating pain, and redressing personal misconceptions Four educational videos for nurses with the aim of improving their knowledge about nursing assessment of cancer pain, the main analgesic groups, opioid analgesics, and under-treated pain Printed educational booklet ‘‘no more pain’’, addressing myths about analgesics and rationale for adhering. Content delivered by nurse over 15 min. Nurse personalized aspects of the booklet to the patient’s pain control regime Educational video, provision of informational booklet. Content explained by nurse who reviewed individual concerns Patient reports on written pain assessment sheet with a summary for professionals highlighting poorly controlled pain Nurse education: four hour lectures with notes on analgesics, misconceptions, and communicating with physician colleagues Provision of information booklet, completion of pain diary, and increased access to a physician who could adjust analgesic
277
278
Table 4 (Continued ) First author
Ward [63]
Ward [64]
Wells [66]
Wilkie [67]
Yates [68]
Yildrim [69]
Comparator
Intervention components
Individually tailored intervention targeting patient attitudinal barriers toward reporting pain and using analgesics. Intervention delivered by telephone followed up with printed mailed copy of educational messages ‘‘RID’’ pain intervention, face to face 20 to 60 min education tackling pain management misconceptions
Usual care
x
x
No improvement in pain outcomes despite reduction in patient barriers questionnaire scores
Booklet about analgesics and misconceptions, follow up telephone call from nurse
x
x
RID pain intervention targeting individual patients or dyad of patient plus carer/significant other 15 min video for patients and carers,
