There is a new paradigm in pain management due to an epidemic of opioid-related adverse events and the need for opioid sparing approaches. Maximum doses of opioid analgesics result in difficult management of post-surgical pain (Jarzyna et al., 2011a). Nurse-specific pain management practices have been studied in relation to medicine; however, the relationship between opioid sparing, nurse-specific, interventions are not clear (Carroll, Atkins, Herold, & Mlcek, 1999; Wu & Raja, 2011). The purpose of this study was to inform a standardized process for pain management in patients undergoing a total-hip or total knee arthroplasty.
Specific Research Aims: Describe the relationship and strength of indicator between nurse-specific pain management variables and pain.
AIM 1: To describe pain management variables and pain (overall and by surgical group).
AIM 2: Examine relationships among variables (overall and by surgical group).
AIM 3: Describe differences in pain between total hip and total knee surgical groups.
AIM 4: Identify predictors of pain among surgical-orthopedic, total hip or total knee, patients receiving services in 4 Community Hospital orthopedic units.
Methods:
The study will included all total hip or total knee arthroplasty discharged from a System of 4 Community Hospital Facilities between March 1, 2016 and April 30, 2017. The sample size was 1639 patients receiving either the total hip or the total knee arthroplasty procedure and discharged from one of the 4 hospitals. The dependent variable (DV) is median pain scores (during hospitalization - patient-report: 0=no pain, 1=mild pain, 2=moderate pain, 3=severe pain) and Acceptable level of pain (met/not met) during hospitalization. The independent variables include demographics (age, sex, BMI, payor), 24-hour morphine equivalent, eGFR<45, RASS or POSS score, average time between nursing pain assessments, acceptable level of pain (met/not met), and the use of adjunctive therapy. Descriptive statistics were generated for all variables. A logistic regression analysis was used to determine the predictive power of the model. A descriptive analysis using descriptive statistics was done for all independent variables The dependent variable (DV) pain scores median (scale 1-10) and acceptable pain level (Met/Not Met).
Analysis of variance was performed for differences between pain levels of total hip and total knee patients. A multiple regression analysis was used to determine the predictive power of the independent variables identified above on the dependent variables in the total hip or total knee patient.
Results:
Correlations: Median pain level during hospitalization (yes/no) significantly correlated with: median pain level day 2 post op, age, bmi, 24hr morphine equiv. during hospitalization, 24hr morphine equiv. day 2 post op, time between nursing pain assess. during hospitalization, and time between nursing pain assess. day 2 post op.
Chi-square: Acceptable pain during hospitalization (yes/no) significant for the following variables: veteran status, ambulation time < 4hrs, RASS time. Also for facility, gender, surgeon, type of surgery (knee/hip), and acceptable pain day 2 post op.
One way Anova: A one-way between-groups anova was conducted to explore the impact of age, BMI, 24hr morphine equivalent during hospitalization, 24hr morphine equivalent day 2 post op, eGFR, time between nursing pain assessment during hospitalization, and time between nursing pain assessment day 2 post op on the levels of acceptable pain during hospitalization. Patients were divided into four groups (no pain, mild pain, moderate pain, and severe pain). There was a statistically significant difference in the amount (mg) of daily morphine equivalent, F (3, 1619) = 4.507, p = .004. Post-hoc comparisons using the Tukey HSD test indicated the mean score for patients in the mild pain group (M = 29.11, SD = 22.75) was significantly different from the moderate pain group (M = 32.88, SD = 23.85). The other groups (no pain, severe pain) did not differ significantly from the mild or moderate pain groups.
Multinomial log regression:
Outcome: Acceptable level of pain (during hospitalization - patient-report) 0=no pain, 1=mild pain, 2=moderate pain, 3=severe pain.
Predictors: veteran status, ambulation in < 4hrs, bmi, daily morphine equivalent during hospitalization, time between pain assessment during hospitalization.
Goodness-of-fit:Pearson value is very high (X2=3409.87) and significant (p=1.0) indicating a good fit of this model.
Conclusion: The strongest predictors of nurse-controlled variables were ambulation in < 4hrs, bmi, and time between pain assessment during hospitalization. The daily morphine equivalent during hospitalization has a relationship to the nurses work, however in an era of opioid addiction, an opioid-sparing approach is preferred. Nurse-controlled variables empowers the nurse to improve patient care while decreasing the patient's risk for post-surgical opioid addictions.
See more of: Research Sessions: Oral Paper & Posters