Despite efforts to improve the management of patients' pain postoperatively with the introduction of guidelines (Wu and Raja, 2011) and the branding of pain as the ‘fifth vital sign’ to raise awareness of the importance of assessment (Campbell, 1996; Morone and Weiner, 2013), many patients continue to suffer high levels of postoperative pain (Rockett et al, 2013; Scher et al, 2018). One reason suggested for this is that postoperative pain is complex, involving both inflammation and nerve injury, which requires a complex person-centred management strategy rather than the administration of analgesia alone (Pogatzki-Zahn et al, 2017). Although most people experience pain after surgery, less than half report adequate pain relief (Institute of Medicine (IOM), 2011). Poorly managed postoperative pain is a global issue (Wu and Raja, 2011). For most patients who report inadequate postoperative pain relief their pain is rated at moderate or higher intensity (IOM, 2011). Prevalence rates of such inadequate pain relief vary by country, from 18% in Norway (Johansen et al, 2012) to as high as 62% reported in South Africa (Murray and Retief, 2016). At discharge, the prevalence rates of severe postoperative pain have been estimated to be in the range of 5-10% (Italy) (Sansone et al, 2015) and 12% (USA) (Buvanendran et al, 2015). In the study by Buvanendran et al (2015), 13% of patients reported severe pain 2 weeks postoperatively. The situation is not that different in the UK, where the prevalence of postoperative pain has been estimated at around 64% (Rockett et al, 2015).
Unrelieved postoperative pain is also associated with increased healthcare costs (Joshi and Ogunnaike, 2005) and delayed discharge (Klopper et al, 2006), as well as negatively affecting patients' psychological and physical recovery (Gillaspie, 2010). It has been demonstrated that in some patients undertreatment of early postoperative pain can lead to the development of chronic pain syndromes. In a prospective study of 30 patients who had had a thoracotomy, 52% still had pain after 1.5 years; a higher level of day 1 postoperative pain was the main predictor for this persistent pain (Katz et al, 1996). Similarly, 58% of 112 patients reported persistent pain 3 months after knee arthroplasty, and those with more severe day 1 postoperative pain reported higher levels of pain at 3 months (Lavand'homme et al, 2014), and worldwide 10-50% of postoperative patients progress to having chronic pain (IOM, 2011). All this highlights the crucial importance of effective management in the acute phase to prevent the development of chronic pain (Meissner et al, 2015).
The first step in providing effective pain management is undertaking objective, accurate and routine pain assessments (British Pain Society (BPS), 2013; Chou et al, 2016). However, pain is subjective, so in large part what the patient is communicating is subject to the healthcare practitioner's interpretation. Interpretation is influenced by the practitioner's values (Fulford et al, 2012) and level of experience (Brant et al 2017), and requires self-awareness to mitigate against the risk of biased judgements (Breivik et al, 2008). In the latest Care Quality Commission (CQC) survey in the UK (CQC, 2019), 8% of patients reported that staff did not do everything to control their pain, and there was a decline, albeit a small one, in those reporting that staff ‘definitely’ did everything they could do to control their pain from 69% in 2017 to 67% in 2018.
Nurses are the health professionals who spend the most time with patients. They usually make the first assessment of a patient's pain and consequently make decisions about analgesia (Melotti et al, 2009). It is therefore important that nurses can assess pain accurately (Klopper et al, 2006).
Patient self-reporting, while open to both practitioner and patient subjectivity, is considered the most accurate measure of pain (Department of Health (DH), 2010). Therefore, this literature review explores nurses' use of patient self-reporting as part of the pain assessment process and the extent to which this might be a contributing factor towards many patients having a poor experience with pain postoperatively.
Design
A structured search was made of the following databases: Cumulative Index of Nursing and Allied Health Literature (CINAHL), Medline Complete, PubMed Complementary Index, PsychINFO and British Nursing Index (BNI). The different stages of the search are outlined in Figure 1 (Moher et al, 2009). The search terms used were: (Nurs* AND patient) AND (pain assess* OR pain examin* OR pain manag*) AND (surg* OR postop*). The initial high volume of results was reduced by applying the following filters to elicit relevant and contemporary research: English language, peer-reviewed and published since 2005. Additional studies were identified using the network method employing the reference lists from studies found in the database search (Timmins and McCabe, 2005). The titles of 574 publications were screened for inclusion.
This search yielded 64 studies, which were then manually searched by abstract for the application of inclusion/exclusion criteria (Table 1). Studies were excluded for the following reasons:
| Inclusion | Exclusion |
|---|---|
|
|
|
No qualitative studies were identified that included contemporaneous data from both patients and nurses from the same clinical location. This left seven primary research studies (see summary of studies in the Appendix) that were deemed relevant and met the inclusion criteria. These were critically appraised utilising Understanding Health Research (Medical Research Council, 2020). All seven studies (Idvall et al, 2005; Sloman et al, 2005; Gunningberg and Idvall 2007; Düzel et al, 2013; Alemdar and Aktas, 2014; Atkinson and Almahdi 2014; Schreiber et al, 2014) had an observational design. While self-selection or allocation bias always needs to be considered when interpreting the findings from observation studies (Howick 2011), these do often take an approach from which conclusions can be drawn about actual clinical practice. A summary of these seven studies is presented in the Appendix.
Results
The literature review identified two key themes:
Nurses underestimation of patients' pain
Across all seven studies nurses consistently underestimated patients' postoperative pain compared with the patients' reporting of pain (Appendix). In the cross-sectional study by Sloman et al (2005), where there were 95 postoperative patients and 95 nurses in four hospitals, participants completed the Short-Form McGill Pain Questionnaire (SF-MPQ) and three visual analogue scales (VAS) to rate overall pain intensity, suffering and satisfaction with pain relief at one point in time. Nurses were found to rate patients' pain significantly lower on all dimension of the SF-MPQ compared with patients. Patients' responses on the SF-MPQ strongly correlated with their responses on the VAS questions, lending additional validity to the pain assessment tools. Nurses completed their pain assessment within minutes of patients' self-assessment, suggesting recall bias was unlikely to explain this lack of agreement.
In another cross-sectional study, the responses of 94 postoperative patients and 47 nurses using the full version of the MPQ were compared (Düzel et al, 2013). Patients' scores were significantly higher than those of nurses on affective, sensory and evaluative sections of the questionnaire, as were the scores for average pain intensity. There was no significant difference in scores for the timing of pain; however, the between-group difference on this subscale and the overall McGill scores were not analysed. Researchers were present during data collection, which could have introduced reporting bias.
A similar methodology comparing nurses' and patients' pain assessments was employed in the study by Idvall et al (2005), but this time using the Strategic and Clinical Quality Indicators in Postoperative Pain Management questionnaire (SCQIPP) and two numerical rating scales (NRS): worst pain during the past 24 hours and satisfaction (Idvall et al, 2005). This was a larger study involving 97 nurses and 286 postoperative patients (general, gynaecological and orthopaedic), but the findings were very similar. Patients rated their worst pain in the preceding 24 hours significantly higher than did the nurses. Nurses did not complete their assessments contemporaneously with patients, hence the reporting in this study may have been affected by recall bias; however, the findings remain consistent with those of Sloman et al (2005) and Düzel et al (2013). Interestingly, patient satisfaction with pain management was rated significantly higher by patients than nurses.
The data collected in one hospital (two surgical departments) in the study by Idvall et al (2005) were further analysed by Gunningberg and Idvall (2007), who reported findings based on responses from 47 nurses and 121 postoperative patients. Patients in general surgery reported significantly higher levels of pain than thoracic surgery patients, and it was in general surgical, not thoracic patients, in whom nurses' significantly underestimated patients' pain levels (Gunningberg and Idvall, 2007). What was similar in both departments, however, was that patients rated their worst pain in the preceding 24 hours significantly higher than was documented in the patient records. For a large proportion of patients (38% general surgical; 23% thoracic) no NRS was documented.
Atkinson and Almahdi (2014) conducted a prospective audit involving 204 medical and surgical patients over 3 non-consecutive days in one hospital. Pain scores could have been documented by any member of the multidisciplinary team but is presumed to be largely the domain of nurses. Of the 176 patients for whom a pain score was documented, this was in agreement with the observed pain score (when asked by the researcher) in 71% of medical patients (n=117), but only 27% of surgical patients. Excluding the results of patients who had no pain at the time of the audit, the documented and observed pain scores only agreed for two medical patients, and for no surgical patients among the 86 patients who reported any degree of pain. Although the authors did not state how much time had elapsed between the documented and observed pain scores (Atkinson and Almahdi, 2014), this association between agreement and size of the pain score dovetailed with the findings of Gunningberg and Idvall (2007). Unsurprisingly, Atkinson and Almahdi (2014) also found that pain management plans relating to the provision of effective analgesia were often inadequate due to the underassessment or poor documentation of many patients' pain.
In another more recent study, a cross-sectional approach was used involving postoperative patients (n=145) and nurses (n=36) at one hospital (Alemdar and Aktas, 2014). Consistent with the other studies described above, patients rated their pain on average higher on the NRS than nurses. However, it was not reported whether or not this difference was statistically significant. The results of a correlation analysis were reported instead, showing no significant correlation between nurses' and patients' ratings of pain using NRS. In contrast, correlation between nurses' and patients' rating of pain on the EQ-5D pain subscale was found to be significant. However, in neither case was the r value reported, which is necessary to interpret the strength of a correlation (Akoglu, 2018). Further, correlation analysis can be misleading, because even high correlation does not mean that the two ratings agree (Bland and Altman, 1986). The extent to which nurses' and patients' ratings of pain differed or agreed in the study by Alemdar and Aktas (2014) therefore remains unknown.
Nurses' knowledge and understanding of pain assessment
Schreiber et al (2014) conducted a quasi-experimental study to determine the effects of an educational intervention on nurses' assessment of pain. The intervention consisted of a 2-day pain management programme. Of 600 nurses invited, 203 agreed to participate; they had between 4 and 10 years' nursing experience between them. The study involved 30 postoperative hip/knee surgery patients pre-intervention and another cohort of 30 post-intervention. Patients were asked to complete pain diaries for 2-3 days postoperatively and their scores were then compared with those documented by nurses. Nurses also completed the Brockopp-Warden Pain Knowledge/Bias Questionnaire (BWPKBQ). Post-intervention, it was found that the mean difference between the nurses' and patients' assessment of pain had reduced (from mean difference of 2, to 1.3, P=0.02), although nurses continued to underestimate pain compared with patients. Despite some nurses exhibiting poor pain management knowledge at baseline, according to the BWPKBQ results, this remained unchanged post-intervention. Likewise, there was no change in pain-related bias, with nurses continuing to spend less time managing pain in patients with ‘non-physical’ problems such as drug misuse or suicide attempt (Schreiber et al, 2014). The overall response rate was poor (23%).
Although the study by Alemdar and Aktas (2014) did not implement an educational intervention, their survey included nurses' level of education. Of their sample of 36 nurses, just over half were educated to associate degree (2 years) or bachelor's degree (4 years) level. The authors suggested that the lack of agreement between the pain assessments of nurses and patients might be related to low educational levels in their sample, although since education was not a focus of this study the suggestion remains speculative. In the study by Düzel et al (2013), a larger proportion of nurses had degrees (just over half of 47 nurses), but educational level did not apparently correlate with incongruent pain assessments. A similar conclusion was reached by Sloman et al (2005). However, these studies were not designed to answer questions related to nurses' level of education and the accurate assessment of patients' pain.
Discussion
This review concludes that nurses' assessment of pain postoperatively often underestimates its severity compared with patients' self-assessment. While patient self-reporting is subjective, it remains the most objective measure of pain (DH, 2010) and the consistent underestimation of severity could explain why postoperative pain is often managed ineffectively (CQC, 2019). Instances of poor pain assessment and management documentation were also identified in this review (Gunningberg and Idvall, 2007; Atkinson and Almahdi, 2014). Dovetailing with research into reduced nurse staffing levels, this lack of documentation suggests that this essential aspect of care is at times being missed (Ball et al, 2018). Accurate and timely documentation is vital for the appropriate continuity of care between shifts and between health professionals; missing documentation risks further mismanagement. Other reasons for ineffective pain management by nurses have been identified: there is evidence, for example, of a need to improve some nurses' knowledge and attitudes towards pain (Brant et al, 2017). These findings are concerning because they highlight failings in one of the fundamental aspects of nursing care. Although this review identified just seven studies, the findings across all of them were strikingly similar, despite the publication dates ranging across 9 years. Further, the studies provided insight into nursing care from five countries across three continents, suggesting that poor pain assessment may be a problem facing many healthcare systems.
Despite the small number of studies identified, some confidence in the findings of this review can be gained because they dovetail with those of similar reviews (Bell, 2000; Solomon, 2001; Bell and Duffy, 2009). Pain assessment has been widely studied, with the consensus that there are discrepancies between nurses and patients. Both Bell (2000) and Solomon (2001) found severe pain to be more underestimated by nurses, which is a similar finding to Gunningberg and Idvall (2007). Nurses are the most researched health profession with regards to the assessment of pain (Solomon, 2001), and it could be interpreted that nurses are the only health professionals to assess patients' pain inaccurately. However, doctors have also often been found to underestimate and disagree with patients' self-reporting of pain (Marquié et al, 2003). Poor pain assessment is therefore a multidisciplinary problem and perhaps, in acknowledgement of that, guidelines stipulate that all health professionals involved in patient care, not solely nurses, have a responsibility to assess pain accurately (Chou et al, 2016).
The studies included in this literature review represent many countries and so provide a wider perspective on the problems with pain assessment. Sweden's healthcare system, like many others, has been under financial and political pressure (Anell, 2005). Further studies conducted in Greece (Chamaidi, 2012) and in Switzerland (van Ransbeeck et al, 2018), which is reputed to be the third best system in the world (Legatum Institute, 2019), also found consistently poor pain assessment, suggesting that it is by no means a problem confined to a single healthcare system or country. Although poor pain assessment might be prevalent across cultures, it needs to be remembered that nurses increasingly assess patients from many different cultures, with the UK as but one example of a country that is ethnically diverse as a result of globalisation (Office for National Statistics, 2012). People from across countries and cultures might express pain in culturally different ways, which can add to the risk of misinterpretation by health professionals. Support for this comes from an Australian study, which found that nurses did not take into consideration the cultural differences of Aboriginal patients, resulting in poor pain assessment and unsafe practice (Fenwick, 2006).
In the studies identified by the present review, the cultural backgrounds of patients and nurses were largely overlooked as factors that might promote or inhibit the management of postoperative pain, suggesting a considerable gap in the literature. In the one study that did consider cultural background, the cultural factors of nurses and patients were documented and found not to be statistically significant between groups (Sloman et al, 2005). However, between-participant differences (ie between any particular patient and the nurse with main caring responsibility for them) were not examined. Further, any one patient may have come into contact with staff from varied cultural backgrounds throughout any one day. Therefore, while Sloman et al (2005) were surprised that cultural factors were not associated with the lack of congruence between patients' and nurses' assessment of pain, a between-participant analysis might have been more appropriate to examine this.
The role of pain education was highlighted by Schreiber et al (2014) with reference to improving nurses' (and other health professionals') assessment of pain. Although the educational intervention they employed did seem to reduce the discrepancy between nurses' and patients' pain ratings, some nurses' knowledge of pain remained poor afterwards. Tellingly, bias remained largely unchanged post-education intervention, with some nurses continuing to spend less time managing the pain of patients with ‘non-physical’ ailments, such as ‘drug abusers’, ‘suicide attempters’, patients who were ‘frequently readmitted’ and ‘confused elderly’ (Schreiber et al, 2014). These findings align with those of Brockopp et al (2003), who found that critical care and medical-surgical nurses and student nurses would apparently devote less time and energy to managing pain in ‘suicide attempters’, ‘substance abusers’ and elderly patients compared with those who had AIDS or cancer.
A study exploring the attitudes of doctors from a variety of specialties found it common for negative attitudes to be expressed with respect to patients who were drug users, and a reluctance to prescribe strong opioids for fear of addiction or misuse (Baldacchino et al, 2010). Reluctance to give patients opioids due to fear of creating dependence was also identified in a study of nurses caring for older patients (Manias, 2012). In addition, this study highlighted that nurses' communication when assessing pain was not always effective, for example not adapting to different levels of communication need (Manias, 2012). A recent review of the literature identified similar communication problems in the context of pain management in patients with dementia (Chandler et al, 2017).
The observational nature of the study by Schreiber et al (2014) means that conclusions cannot be drawn as to cause and effect of the educational intervention. However, it might be that the method of delivery of the package, largely didactic, in itself had limited effectiveness. A more recent study of critical care nurses (Lewis et al, 2015) suggested that bias towards vulnerable groups such as drug users might be improved by implementing educational interventions in small groups. Future studies need to focus more on active learning methods that appear to facilitate deeper learning (Hew and Lo, 2018), and a randomised controlled design is needed to determine effectiveness.
A root problem of nurses' lack of knowledge is that the topic of pain appears to be neglected at undergraduate level. A survey of undergraduate healthcare curricula, including medicine, physiotherapy and nursing, found pain education in the UK to be inadequate, with only 12 hours on average of pain content in the programmes surveyed (Briggs et al, 2011). The findings from a more recent exploration of pain education in undergraduate nursing curricula in the UK would suggest that the situation is little improved, with the word ‘pain’ hardly featuring in programme documents and web resources (MacIntosh-Franklin, 2017). Such inadequate provision of knowledge and understanding at undergraduate level would seem to indicate that difficulties in improving nurses' pain knowledge will persist.
Adequate pain management necessitates an interprofessional team approach, so one suggested advance is for an interprofessional approach to pain education (Gordon et al, 2018). Interprofessional education (IPE) can be difficult to operationalise due to the logistics of managing many students and staff across multiple curricula (Gordon et al, 2018). However, a recent IPE study found an e-learning package to be effective for students' learning across seven healthcare programmes, with positive student evaluations (Watt-Watson et al, 2019), suggesting that many logistical barriers might be overcome through the use of technology. Whichever way it is delivered, at undergraduate or postgraduate level, pain must be given sufficient prominence, and any pain education needs to effectively target prejudiced practices to avoid unnecessary suffering by vulnerable groups of patients.
The challenge of improving pain assessment through education appears within the wider literature too. Ene et al (2008), who implemented a 6-month educational programme for nurses and investigated the impact 2 years later, found no significant improvement in nurses' ability to assess patients' pain accurately, which was also the finding in reviews by McCaffery et al (2000) and Bell and Duffy (2009). This is a significant challenge, especially given the findings of a report by the Royal College of Anaesthetists that nurses' inadequate knowledge contributed to the failure to meet patient satisfaction in pain management (Rockett et al, 2015). Standards now stipulate that pain management services must demonstrate engagement with training and education in accordance with best practice (McGhie, 2015). Achieving this necessitates ongoing education on pain management for all relevant health professionals (American Society of Anaethesiologists (ANA), 2012). However, due to cost pressures in healthcare systems, educational programmes are not always implemented and nurses are often not afforded the time or resources to attend (Iacobucci, 2017). Without sufficient support for postgraduate learning, it is unlikely that there will be improvements in poor pain management.
Inaccurate pain assessment leads to poorer treatment of pain, resulting in patients receiving inadequate analgesia (Sloman et al, 2005). UK standards recommend that patients in significant pain must be treated within 30 minutes and reassessed after an appropriate amount of time (Rockett et al, 2015). If this does not happen, the result is preventable suffering, which violates one of the ethical tenets of healthcare practice: to do no harm. Unrelieved pain can considerably affect a patient's recovery because it reduces mobilisation (Strassels et al, 2004), which in turn increases the risk of complications such as venous thrombosis (Chung and Lui, 2003). The repercussions of this can be unnecessary distress and anxiety, and delayed wound healing time (Pinto et al, 2016). This also leads to increased length of stay in care settings with associated excess costs (Shang and Gan, 2003).
Although not solely due to postoperative pain, delayed discharges are reported to cost the NHS £100 million per year, 1.2 million bed days and many cancellations of elective operations (Rojas-García et al, 2018). Poor pain assessment is an avoidable contributing factor to excess socioeconomic costs. Pain management in hospitals can be particularly challenging due to the shift from inpatient to day-case surgery (Breivik et al, 2006) placing time pressures on management decisions. Time pressures are problematic in the context of patients with communication difficulties, extending patient encounters; this can be more common with older patients (Manias, 2012). An ageing population also means an increase in the number of elderly surgical patients with severe medical comorbidities, adding to the complexity of postoperative pain management (Rockett et al, 2015). Acute pain services exist to provide specialist multidisciplinary support to some extent throughout the NHS (Rockett et al, 2015). However, provision varies, with most providers surveyed between 2014 and 2016 not offering the service overnight or at weekends (Rockett et al, 2017). Therefore, timely specialist input may not be available for some patients.
In line with the Nursing and Midwifery Council's Code (2018), nurses should always respond to the physical and psychological needs of their patients. If nurses do not accurately assess pain, nor deliver evidence-based pain management, they are failing to uphold their duty of care. For example, Grinstein-Cohen et al (2009) and Manias (2012) found nurses were fearful of overmedicating patients with opioids due to the risk of addiction. It may be that some nurses feel they are upholding their duty of care by reducing the chances of patients developing a drug addiction, further supporting the need for effective education in pain assessment and management. To address such education needs, a hospital in Australia created the ‘pain resource nurse’ role, whose remit is to facilitate evidence-based change across the hospital (Allen et al, 2018). Eight years after its implementation, while documentation of pain assessments had significantly improved, a survey of the pain resource nurses identified that a knowledge gap remained in respect of the risk of inducing opioid dependence (Allen et al, 2018); this underscores the need for ongoing training and education. Qualitative studies are needed to explore why such pain beliefs might persist, in both nurses and patients, and elucidate what might help change these.
There were some important methodological weaknesses in the studies identified by this review. No studies described the pain assessment approach used by nurses. This limits confidence in the findings and the ability to replicate studies or make comparisons between them. Some studies did not analyse data for between-group differences, ie between nurses and patients; instead, they analysed for correlations, which are not the same as assessing for agreement or statistically significant differences between groups. Further, only Düzel et al (2013) used (in one instance) a reliability statistic such as the kappa coefficient, which gives a more accurate determination of agreement between ratings on scales such as the NRS (Sim and Wright, 2005). Finally, all studies required the nurses to rate patients' pain, which is of questionable validity when pain is subjective. This nurse assessment does not reflect actual clinical practice, which involves eliciting patients' self-reporting of their pain. However, these studies do provide powerful insight into the vulnerability of patients' self-reporting being misinterpreted by health professionals, as confirmed by the audit of patient records by Atkinson and Almadi (2014). This study also highlighted the need for ongoing audit, not only to identify deficiencies but to evidence improvement (or otherwise) in response to any interventions implemented to correct deficiencies.
To accurately assess pain, guidelines (BPS, 2013) suggest a multidimensional approach, whereby associated factors are taken into consideration as part of assessment. The biopsychosocial theory of pain suggests that the experience of pain is best understood when viewed in relation to biological, psychological and social factors (Wright, 2014). This highlights that pain cannot be treated effectively using a single approach. Standards recommend that all inpatients with acute pain should have individualised treatment plans based on the principles of multimodal pharmacological and non-pharmacological treatments (Rockett, 2015). In acknowledgement of this, multidimensional approaches to pain management are a feature of enhanced recovery after surgery (ERAS) guidelines (Ibrahim et al, 2013; Feldheiser et al, 2016). A driver for ERAS has been to reduce length of hospital stay, and therefore healthcare costs, for patients after surgery (Ibrahim et al, 2013). The accelerated patient turnover has put pressure on the amount of time nurses can spend with each patient, which is compounded by reduced nurse staffing levels (Griffiths et al, 2016). Across many countries low nurse staffing levels have been associated with necessary nursing care being missed, including pain management (Ball et al, 2018). In a systematic review of patients' experiences of ERAS (Sibbern et al, 2017), a common theme reported by patients was inadequate time to absorb information and ask questions at their pre-admission appointment. A randomised controlled trial (RCT) of 760 patients, who either received written information or not, found that providing written preoperative information improved patients' knowledge about postoperative complications related to severe pain, but did not allay their beliefs about pain management, such as concerns about opioid addiction (van Dijk et al, 2017). Patients need to be given sufficient time for discussion, as well as for the provision of information: for example, Bayman et al (2019) reported that patients with expectations of high levels of pain preoperatively are at increased risk of moderate-severe pain postoperatively. However, there is also evidence that many nurses hold erroneous views about the risk of postoperative opioid addiction (Manias, 2012; van Dijk et al, 2017), which can only serve to reinforce patients' pre-existing erroneous beliefs. In the cross-sectional component of the study by van Dijk et al (2017), which included 1184 nurses, the findings showed that the better the pain education of nurses, the higher their knowledge and, crucially, the more positive their pain-management beliefs. This again reinforces that ongoing pain education is crucial for improving postoperative pain management.
A systematic review by Sibbern et al (2016) identified another important theme, namely that patients sometimes thought that nurses were too rigid in adhering to a standardised protocol, which was seen as a barrier to person-centred care. Some of the studies identified by this present review did use the MPQ to facilitate a multidimensional assessment (Sloman et al, 2005; Düzel et al, 2013). Yet the findings from these studies were no different to those of Alemdar and Aktas (2014), who simply used a numerical pain scoring tool, a one-dimensional approach to assessing pain (Morone and Weiner, 2013). This only serves to emphasise that a pain assessment tool is only as good as the decisions made by the health professional using it.
Conclusion
This literature review has identified that the incongruence between nurses' and patients' assessment of postoperative pain remains a problem internationally. It suggests that, worldwide, patients are suffering from avoidably high levels of pain, presenting a strong ethical challenge to healthcare systems and the health professions, notably nurses. Pain management knowledge gaps have been identified in some nurses, but so far educational interventions seem to be of modest benefit only.
Biased attitudes have also been identified in some nurses towards the management of pain in some vulnerable patients groups, which also need to be addressed if poor care is to be avoided. Future RCTs are needed to determine the effectiveness of different educational interventions at improving nurses' knowledge about, and attitudes towards, effective assessment and management of patients' pain postoperatively, as well as other members' of the multidisciplinary team. Key to supporting any educational interventions is a greater understanding of nurses' and other health professionals' pain beliefs, as well as those of patients, with clarity about what is needed to change those beliefs. If improvements are to be realised, there must be a greater emphasis on pain education across all undergraduate healthcare disciplines.