Improvements in pain outcomes in a Canadian pediatric teaching hospital following implementation of a multifaceted , knowledge translation initiative

1Department of Pharmacy, Sunnybrook Health Sciences Centre; 2Doctor of Pharmacy Program, Leslie Dan Faculty of Pharmacy, University of Toronto; 3Child Health Evaluative Sciences, The Hospital for Sick Children; 4The Hospital for Sick Children; 5Department of Haematology/ Oncology, The Hospital for Sick Children; 6Graduate Department of Pharmaceutical Sciences, Leslie Dan Faculty of Pharmacy, University of Toronto, Toronto, Ontario; 7Department of Pharmacy and Herzl Family Medicine Centre, Jewish General Hospital, Montreal, Quebec; 8Service des Urgences pédiatriques, Hôpital Armand Trousseau, AP-HP, Paris, France; 9INSERM, US953, UPMC Paris 06; 10Department of Anesthesia and Pain Medicine, The Hospital for Sick Children, University of Toronto; 11Clincal Social and Administrative Pharmacy, Leslie Dan Faculty of Pharmacy, University of Toronto, Toronto, Ontario Correspondence: Dr Anna Taddio, Leslie Dan Faculty of Pharmacy, University of Toronto, 144 College Street, Toronto, Ontario M5S 3M2. Telephone 416-978-8822, fax 416-978-1833, e-mail anna.taddio@utoronto.ca Pain control is considered to be a fundamental human right (1). In addition to the obvious humane reasons for assessing and treating pain, there is evidence that undertreatment of pain can lead to persistent pain, alterations in nociceptive processing (2,3), and emotional and psychological complications (4). Well-managed pain, on the other hand, is associated with faster recoveries, fewer complications and decreased use of health care resources (5,6). Presently, there is a plethora of evidence-based pain assessment and management techniques to assist in reducing the burden of pain among hospitalized children. Despite this, several epidemiological studies have reported a high prevalence of pain (49% to 87%) in hospitalized children, with more than one-half suffering from poorly managed pain (7-9). As a orIgInal artICle


The knowledge gap
In 2004, we conducted an audit of pain practices at a large tertiary/ quaternary pediatric hospital in Toronto, Ontario.This was a one-day audit of all inpatients including patients in medical, surgical and intensive care units (ICUs).A structured questionnaire was verbally administered to collect patient demographics and determine pain prevalence, pain intensity and the type of analgesics received during admission and in the previous 24 h.Charts were reviewed to establish the frequency of documented pain assessments in the previous 24 h and to verify analgesic administration.This study (9) revealed suboptimal assessment and treatment of children's pain.The results of our 2004 audit highlighted gaps between pain assessment and management recommendations and routine clinical practice, and led to the development and implementation of multifaceted knowledge translation (KT) initiatives aimed at improving pain practices.The multifaceted KT interventions included the use of education, reminders, and audit and feedback (Table 1), all of which have evidence supporting their use (16).
The purpose of the present study was to evaluate the global impact of KT initiatives on pain outcomes, including pain processes (eg, pain assessment documentation and pain management practices, including pharmacological, physical and psychological interventions) and clinical pain outcomes (prevalence and intensity of pain), by comparing our findings with those of our previous audit; and to benchmark further pain practices, particularly with respect to procedural pain assessment and management, and opioid utilization patterns.

Settings, patients and procedures
The present cross-sectional observational study included all inpatients at The Hospital for Sick Children (Toronto, Ontario) on the selected study day.The study was approved by the hospital's Research Ethics Board.Because the study did not influence the care of patients, consent for participation was waived by the Research Ethics Board.
The study day was a randomly selected weekday in September, 2007.Hospital staff were notified of an upcoming pain audit by e-mail and posted signs in the nursing units.To avoid a change in pain practices for the day of the audit, messages had limited information regarding the purpose of the study and did not specify the exact date of the audit (instead, a range of possible dates was given).
All inpatients younger than 18 years of age whose names appeared on the 08:00 hospital census the day of the audit were included in the study.Data were collected from patient medical records and recorded on a standardized data collection form by either health care professionals and/or trainees working in the hospital.Data collectors received a copy of the protocol and data collection form before the study and participated in a 2 h training session on the morning of the audit.The data collectors were assigned to different units within the hospital, where they collected data for all the inpatients in those units.Study investigators participated in the data collection, and were available to answer questions and provide clarification as required.

Demographics
Demographic variables included age, sex, type of service (eg, medical, surgical, ICU) and whether the child was intubated at the time of the audit.

Pain assessment and management
As per the previous audit ( 9), information was collected on the frequency of pain assessment documentation, type of pain assessment tool used, pain intensity rating and pain management interventions for the preceding 24 h.Pain management interventions included the following: pharmacological (simple analgesics [eg, acetaminophen, nonsteroidal anti-inflammatory drugs (NSAIDs)], opioids [eg, codeine, morphine, hydromorphone and fentanyl] and adjuvant agents [eg, gabapentin, ketamine, amitriptyline and clonidine); physical (eg, heat or cold therapy); and psychological (eg, distraction) strategies.Finally, the type and number of painful procedures and procedural pain assessment and management practices were documented during the same 24 h period.Figure 1 illustrates the flow of data that were collected in the study.

Study outcomes
Rates of documented pain assessments were determined during usual care and at the time of procedures.Pain intensity scores obtained from individual pain tools were converted to a common four-level metric (none, mild, moderate and severe pain) to summarize results.The scores for the Verbal Descriptive Scale were converted into these four levels using the cut-off points validated by Jones et al (17) whereby no pain was converted to none; a little pain to mild; medium pain to moderate; and a lot of pain to severe.Scores on the numerical rating scale (NRS) were also converted as follows: 0 = no pain, 1 to 3 = mild pain, 4 to 6 = moderate pain and >6 = severe pain (17).The Face, Legs, Arms, Cry and Consolability scale was converted using the same cut-off points as the NRS.For the Premature Infant Pain Profile, scores of 0 to 6 were converted to none, scores of 7 to 12 to moderate pain, and scores >12 to severe pain.Pain assessed without a documented tool (eg, physiological parameters) or by an unspecified tool were not included.The number of procedures, defined as any medical, nursing, surgical, diagnostic or therapeutic activity, were tabulated with a specific focus on painful procedures.Using the definition proposed by Carbajal et al (18), a procedure was considered to be painful if it invaded the child's bodily integrity, causing skin or mucosal injury by the introduction or removal of foreign material into the airway, or digestive or urinary tract.Additional procedures were considered to be painful, as designated in previous studies (eg, adhesive tape removal) (18)(19)(20)(21).Finally, the percentage of children with documented pharmacological, physical and psychological interventions during usual care and at the time of procedures were calculated.

Data analysis
Data were analyzed using Excel 2010 (Microsoft Corporation, USA).Descriptive statistics were used to summarize data, including means and SDs, medians and interquartile ranges (IQRs) for continuous data, and frequencies and percentages for categorical data.Comparisons in the median number of procedures among children under the care of different medical services (ie, medical, surgical, intensive care) were performed using the Kruskal-Wallis H test. Post hoc comparisons were performed using the Mann-Whitney U test.c 2 tests were used to compare percentages between the current and previous audit (9) in pain assessment documentation, pain management, pain prevalence and pain intensity.P<0.05 was considered to be statistically significant.

Demographics
A total of 265 children were included.The median age was 4.2 years (IQR 0.3 to 12.0, mean [± SD] age 6.1±6.2 years).The proportion of children admitted to each service type is summarized in Table 2.

Pain assessment
In total, 167 (63%) of the 265 children had at least one documented pain assessment in the previous 24 h (Figure 1).Thirty per cent of children had one to two assessments, 17% had three to five assessments, and 16% had six or more assessments in the previous 24 h.There were 666 pain assessments documented for these 167 children, translating to a median of three (IQR 1 to 6) assessments per child.
Children on surgical units (78%) and ICU (78%) were more likely to have a documented pain assessment compared with patients on medical units (46%; P<0.01 for both analyses).The pain assessment tools and the ages of children in whom they were used are presented in Table 3.The NRS was the most frequently used tool to assess pain.

Procedural prevalence and pain assessment
A total of 154 (58%) children underwent 783 procedures (Figure 1).The most commonly performed painful and nonpainful procedures are summarized in Table 4, which account for 95% of all procedures.Forty-nine per cent of children had at least one painful procedure documented (mean 3.1±3.3per child, median 2, range 1 to 20, IQR 1 to 4).Twenty-eight children (21.4%) had a documented pain assessment at the time of the procedure.Overall, few painful procedures were accompanied by a documented pain assessment at the time of the procedure (Figure 1).
A greater proportion of children underwent painful procedures in the ICU (81.8%) when compared with children in medical units (51.2%) or surgical units (27%), P<0.001 for both comparisons.Children in the ICU had an average of 4.9±4.0(median 4, range 1 to 15, IQR 2 to 7) painful procedures per day compared with 1.7±1.2(median 1, range 1 to 4, IQR 1 to 2) for children in surgical units and 2.3±2.8(median 1.5, range 1 to 20, IQR 1 to 2.8) for children in medical units (P<0.001 for both analyses).There was no significant difference between surgical and medical units (P=0.14).

Pain prevalence
Of the 167 children with at least one documented pain assessment, 66 (44%) experienced some pain in the preceding 24 h (Table 5).Patients on surgical units were more likely to experience pain compared with those on medical units (56% versus 29%; P<0.01).Of the 40 painful procedures accompanied by a documented pain assessment, 12 (29.3%)were rated as painful.

Pain intensity
For 89% of the 167 children with at least one documented pain assessment, pain severity could be summarized using the categories of none, mild, moderate and severe (Table 5).More than two-thirds of patients experienced no pain or only mild pain in the previous 24 h.

Pain management
Of the 265 children, 156 (58.9%) received at least one documented pain management intervention (Table 6).Patients on surgical units (73%) and in the ICU (84%) were more likely to receive a painrelieving intervention compared with patients on medical units (37%; P<0.01 for both analyses).
In the 66 children with documented pain (either mild, moderate or severe), 55 (83.3%) received a pain management intervention.Fiftyfour (81.8%) had documentation of at least one pharmacological intervention (analgesic and/or adjuvant), six (9.1%) received a physical intervention and five (7.6%) received a psychological intervention (Table 7).
The prevalence of opioid use according to pain intensity was as follows: 100% of the 13 children with severe pain, 64.7% of the 34 children with moderate pain and 57.9% of the 19 children with mild pain.
For the 407 painful procedures performed in 131 children, 51 (12.5%) were accompanied by a pain management intervention at the time of the procedure.A pain management intervention was administered at sometime within the 24 h period for 284 (69.8%) of procedures (Table 4).

Comparison with previous audit
Table 5 presents a comparison between the previous and present audit results.The frequency of pain assessment documentation was significantly higher, the prevalence of pain was significantly lower, and the proportion of children with severe pain was significantly lower in the present audit.A significantly higher proportion of patients experiencing severe pain received an opioid in the preceding 24 h in the present audit.

DISCUSSION
While previous studies have demonstrated that various quality improvement interventions improve process outcomes (eg, pain assessment documentation, analgesic administration) (22)(23)(24)(25)(26), to our knowledge, the present study was the first to also show improvements in clinical outcomes (eg, pain intensity scores) for patients following implementation of multifaceted KT interventions.Our audit of pain assessment and pain management practices revealed that there were significantly higher rates of pain assessments.Pain management interventions were also more frequently administered, with the use of multimodal analgesia and continuous opioid infusions.Fewer children experienced pain, and pain intensity was significantly lower.More specifically, we found that pain was assessed in approximately two-thirds of children.This represents a more than twofold increase since the implementation of a targeted pain KT dissemination plan across the hospital.However, only one-fifth of children had pain assessed at the time of painful procedures.One-half of the children experienced at least one painful procedure, with an average of three per child per day.Pain prevalence was lower in the present audit when compared with our initial audit, with slightly less than one-half of children with a documented assessment experiencing pain in the preceding 24 h.In addition, significantly fewer children in the present audit experienced severe pain.More than one-half of children received a documented pain management intervention, with pharmacological interventions being the most common.All patients who experienced severe pain received an opioid, which also represents an improvement from the first audit.
Interestingly, approximately one-third of all children received an opioid in the preceding 24 h, with morphine being the most common (16%), followed by codeine.Surprisingly, approximately one-fifth of these patients had no documented pain assessment during this time frame.It is important that pain be reassessed soon after any pharmacological intervention to guide further interventions and ensure pain relief goals are achieved.Of the patients who received an opioid, more than one-half were given concomitant acetaminophen, with or without an NSAID, and approximately 10% were given adjuvant therapy, indicating use of multimodal analgesia practice.This is in keeping with the WHO's standard of providing multimodal analgesia and reflects an improvement over the first audit, in which analgesic interventions consisted primarily of single agents.Another positive finding is that almost one-half of opioids were administered by continuous infusion.This suggests that children in pain were being administered regularly scheduled analgesia, which is preferable to reactive, symptom-triggered administration.
The use of physical and psychological measures to reduce perception of pain and enhance comfort was highlighted in the hospital Pain Management Clinical Practice Guideline and staff education was provided on these strategies (Table 1).Despite this, only 15% of children were recorded as receiving a physical or psychological intervention.Previous studies have also reported that nonpharmacological pain-relieving strategies are infrequently used (27)(28)(29).Because the present study involved an audit of patient medical records, it was not possible to determine whether nonpharmacological measures were not performed, or whether they were performed but not recorded.It is, however, important to acknowledge that nurses' and other health care providers' documentation should reflect all care provided to patients.
It was disappointing that only a small proportion of children received a pain management intervention at the time of a painful procedure, despite considerable evidence of effective procedural pain management strategies (12,13).It is somewhat reassuring, however,  that more than two-thirds had received an intervention in the preceding 24 h.Notably, procedures were considered to be painful according to the definition proposed by Carbajal et al ( 18), yet from the subset of painful procedures that were accompanied by a documented pain assessment, only one-third were actually rated as being painful.Given these findings, the definition of painful procedures and apparent lack of procedural assessment and analgesia administration warrants further investigation.Specifically, the added benefit of breakthrough analgesia in the presence of ongoing analgesia needs to be documented.Despite the improvements in pain processes and patient outcomes, there remains room for improvement.One-third of children had no documented pain assessment and pain assessments were infrequently documented during painful procedures.One-fifth of children who were given opioids did not have documentation of pain assessment.Few patients had either physical or psychological interventions documented.
Our findings have informed the development and implementation of additional KT initiatives (eg, revised and expanded educational sessions and materials).Development of an online interprofessional pain curriculum and a five-week interprofessional pain clinical placement education program for health care trainees is also underway.Numerous reminders of pain resources at the hospital exist (posters and e-mails), and the results of this audit are being disseminated as a further 'audit and feedback' intervention.In addition, audits are being planned for the future.

Study limitations
Interpretation of these results should acknowledge several limitations.First, data were collected by a retrospective review of patient medical records, and documentation may have been incomplete and/or variable.Together, these factors limit our ability to extrapolate the results to actual practices.Nurses and other health care providers may be performing unrecorded pain assessments and providing undocumented pain-relieving (physical or psychological) interventions.Second, some children were excluded from the analysis of pain prevalence and pain intensity because pain was assessed using inappropriate or less reliable methods such as physiological parameters; however, this only represents a small percentage (<10%) of all assessments.Overall, the data suggest that the majority of pain assessments were performed using an appropriate tool.In addition, caution should be taken when interpreting comparisons between the present and previous audits because there were some methodological differences.In the previous audit (9), children and/or caregivers were asked during an interview to recall pain in the previous 24 h, whereas in the present audit, this was determined from documented pain scores in the medical chart.It is reassuring to note that other studies have shown strong associations between recall and actual average pain ratings for the same time frame (30); thus, this was considered to be an appropriate comparison.When interpreting the relative effectiveness of the KT initiatives, it is important to note that the impact of audit and feedback is likely to be greater when baseline adherence to recommended practice is low and when feedback is delivered more intensively (31).The nonexperimental nature of the study prevents us from concluding that the changes were definitely caused by our KT interventions rather than other factors such as changes due to time.However, we believe it is very unlikely that the observed improvements could have been achieved without our KT interventions.Finally, it is not possible to evaluate the impact of the individual components of the multifaceted KT intervention on the pain outcomes.The combined strategies likely contributed to the overall positive impact.

CONCLUSIONS
After implementation of a multifaceted KT intervention, we observed improvements in both process and clinical outcomes including significantly higher rates of documented pain assessment and pain management interventions, and significantly lower pain prevalence, with fewer children experiencing severe pain.Although these results are encouraging, there remains room for improvement.This information has been used to revise existing KT initiatives and develop new ones to support the adoption of best practices, and future audits are planned to assess their effectiveness.

Table 1 Knowledge translation (KT) initiatives at The Hospital for Sick Children (Toronto, Ontario) since 2004 KT initiative and target Objective example Timeline education
nursing admission assessment form; and links to pdfs of the evidence-based pain intensity tools recommended for use at our organization Pain Management Clinical Practice Guideline -development and roll-out; key components of the clinical practice guideline include: developing a pain-goal collaboratively with child and family; link to the Pain Assessment Policy; and guidance on the '3 Ps' of pain control (ie, pharmacological, physical and psychological strategies) Development of Pain Resource Centre (PRC) on www.AboutKidsHealth.ca2004 Reminders Nurses Improve pain assessment documentation Relocated pain assessment documentation to appear more prominently on nursing flow-sheet and inclusion of check boxes to select individualized pain intensity tools 2005 Doctors Improve analgesic prescribing Developed option for prescribing simple analgesics (eg, acetaminophen and ibuprofen) by selecting a tick box, which automatically orders appropriate medication dose and schedule 2005 Health professionals Improve pain practices E-mail reminder from hospital executive to use Pain resources (eg, Pain Assessment Policy and Pain Management Clinical Practice Guideline) 2005 audit and feedback All professional groups Improve pain practices Widespread dissemination of first audit results, including feedback of unit-specific results to each unit 2004, ongoing

Table 4 Documented pain management interventions administered to recipients of procedures (n=783), according to procedure type Procedure type Procedures Pain management intervention administered to recipients of procedure at time of procedure anytime in prev 24 h any pharmacological any physical any psychological any intervention any pharmacological Painful*
(21) presented as n (%).*Procedures were considered painful according to the definition proposed by Carbajal et al(18); † Pain severity of different procedures was according to categorization by Stevens et al(21).Note: table does not include procedure types that accounted for less than 1% of the 783 procedures.PICC Peripherally inserted central catheter; prev Previous