Pain Management Protocol in Abdominal Surgeries in City of Erbil

The pharmacological regimen includes non-opioid analgesics as the foundation of pain control. Patients will receive either ibuprofen 400 mg PO q6h or diclofenac 50 mg PO q8h for the first 48 hours postoperatively, unless contraindicated. This will be complemented by acetaminophen 1,000 mg PO q6h, not exceeding 4,000 mg/day. For breakthrough pain, tramadol 50-100 mg PO q6h is administered on an as-needed basis, with its use limited to the shortest duration necessary to manage acute pain. Adjuvant medications include gabapentin, continued at 100-300 mg PO TID and titrated based on patient response and tolerability, and cyclobenzaprine 5 mg PO TID as needed for muscle spasms.

Discharge phase This phase is designed to investigate the impact of the continuity of care provided from the previous phases and monitor progress following discharge. Prior to discharge, we will set a training session for each patient to illustrate the proper use of pharmacologic and non-pharmacological interventions at home. Furthermore, we will conduct a phone call 48 hours post-discharge to assess pain control and overall recovery and address any concerns or complications that may have arisen.

Randomization Eligible patients will be randomly assigned to the intervention group and control group in a 1:1 allocation ratio using computer-assisted random number generation. To account for the multicenter design, stratified randomization will be adopted based on hospitals. Permuted block sequences with variable block sizes of 2,4,6 were implemented to ensure random allocation to the study arms. The randomization list and randomization cards were produced [see Appendix 3] and will be kept in sealed, opaque envelopes to ensure allocation concealment. These envelopes will be opened by the treating physician only after patient eligibility is confirmed and informed consent is obtained. The randomization process was processed using "blockrand" package in R.

Outcomes:

The primary outcome of the current study will be the change in pain intensity assessed using a visual analog scale (VAS) comprised of 11 points, where zero represents no pain, and 10 represents the worst imaginable pain. The assessment of NRS will be conducted at 2, 6, 12, 24, 48, and 72 hours postoperatively. Based on the individual NRS scores vs. time profile, the area under the curve over the time frame of 72 will be calculated for each patient.

Secondary outcomes will include recovery time, patient satisfaction, opioid consumption, functional recovery, and adverse effects. Recovery time will be defined as the maximum of the following postoperative milestones: time to first ambulation, first oral intake, first flatus, or length of hospital stay. Patients will be assessed for the previous milestones daily from the day of surgery until discharge. Functional recovery will be evaluated using the Quality of Recovery-15 (QoR-15) questionnaire [see appendix 2] administered preoperatively and on day 7 postoperatively. Patients' satisfaction will be evaluated on a 5-point Likert scale from 1 (very dissatisfied) to 5 (very satisfied). Opioid consumption will be measured as the cumulative opioid doses administered at 24, 48, and 72 hours postoperatively in oral morphine milligram equivalents (MME). The incidence of each possible opioid-related adverse effect, including nausea, vomiting, respiratory depression, pruritus, and urinary retention, will be assessed at the time of discharge or postoperative day 7, whichever comes first.

Data collection Data collection will be adopted through a standardized data collection form [See appendix 4] to minimize any inconsistencies in data collection between the two contributing centers in the current study. Furthermore, dedicated research coordinators who will be responsible for the data collection will receive brief training on how to fill out this form. We designed the form to adequately capture the baseline patient characteristics, measures of exposure, potential confounders, and primary/secondary outcome measures and to align with the current recommendations for conducting clinical trials assessing pain management following the surgery. A recent international consensus identified six key standardized endpoints that should be included in clinical trials to evaluate interventions designed to improve patient comfort after surgery. These were: pain intensity (at rest and during movement) at 24 h after surgery, postoperative nausea and vomiting (0-6 h, 6-24 h, overall), one of two quality of recovery scales (QoR Score or QoR-15), time to gastro-intestinal recovery, time to mobilization, and sleep quality1 Items included in the data collection include baseline demographic and clinical data attained during preoperative visits and intraoperative data recorded by anesthesiologists. Opioid consumption and VAS scores recorded at different time points will be extracted from the patient's medical record. Similarly, recovery milestones (first ambulation, oral intake, flatus, and readiness for discharge) recorded for each patient will be verified before extraction into our form. Functional recovery was assessed using the Quality of Recovery-15 (QoR-15) questionnaire preoperatively and day 7 postoperatively. We will determine the incidence of different adverse effects on a daily basis, starting from immediately postoperative till the day postoperative or discharge, whichever comes first. After discharge, the total length of stay will be calculated, and any complications recorded during the 2-week follow-up following the discharge will be recorded.

Sample size calculation:

The principle of the sample size calculation for the current study includes accounting for the main effects of intervention, design effects for clustering, and adjustment of intra-cluster correlations given pre-post pain assessments. The first step in the sample size calculation includes the calculation of the expected sample size in a basic study design using the following formula:

n_"initial" =(2(Z_(α/2)+Z_β )^2⋅SD^2)/Δ^2 Where (α) is the desired significance level, (1-β) represents the power, (Δ) is the expected difference between groups, and (SD) is the pooled standard deviation of the outcome measure. Based on a previous study2, the minimal clinically important difference (MCID) in VAS scores is 1 point with a 2-point standard deviation. Considering a priori α of 0.05, the initial sample size is calculated as 99 patients.

The design effect (DE) accounting for the clustering of participants from multiple centers will be then calculated using the following formula:

"DE"=1+(m-1)×"ICC" Where (m) is the average number of participants per cluster, assumed to be 99 from the first step of the analysis, and ICC is the intra-cluster correlation coefficient parameterizing the correlation between pain scores within each cluster. The ICC is assumed to be 0.05, suggesting little between-hospital variability. This assumption is based on the implementation of a standardized protocol for pain management across all the included centers, as well as an approximate estimate from a previous study3.

Accordingly, the adjusted sample size for multi-center clustering will be:

n_adjusted=n_"initial" ×"DE" The adjusted sample size is then calculated as 257 patients.

Considering the adjustment of pre-post correlations within individuals in the same cluster, as previously recommended by Teerenstra et al.4, the sample size calculation formula was then modified to be:

n_"eff" =n_"adjusted" ×(1-r^2)

Where r is the correlation between baseline and follow-up cluster means and calculated using the following formula:

r=(n×〖"ICC" 〗_c+〖"ICC" 〗_s×(1-"ICC" ))/(n×"ICC" +1) Where n is the number of participants per cluster, ICC is the intra-cluster correlation coefficient, ICCc is the cluster autocorrelation (correlation of cluster means between baseline and follow-up), and ICCs is the subject autocorrelation (correlation of individual subject scores between baseline and follow-up). We assumed an ICCc of 0.4 (moderate correlations) based on the implementation of a standardized intervention protocol across the included hospitals. Furthermore, we assumed ICCs of 0.5 since individual responses at different time points are essentially correlated due to origin from the same patient. The final calculated sample size is 76 patients, accounting for the 4 components: intervention effects, design effects, within-class correlations between pre-post measurements, and potential dropout. A sensitivity analysis was further conducted to investigate our assumptions regarding ICCs and ICCc, suggesting a maximum of 120 patients are calculated from different ICC estimates [See Appendix 6]. To account for potential dropouts, we additionally included 15% over the calculated size. Therefore, our final recommended sample size is 132 patients allocated in a 1:1 ratio between the two hospitals. Within each hospital, 33 patients will be equivalently distributed to either control or intervention groups. The full R code utilized for this calculation and sensitivity analysis is provided in Appendix 5.

Statistical analysis plan After the normality check, continuous variables will be summarized as means and standard deviations for normally distributed data or medians and interquartile ranges for non-normally distributed data. Normality testing will be adopted using the Shapiro-Wilk test. Summaries for categorical variables will be presented as frequency and percentages (n, %). The median VAS scores within each cluster will be compared longitudinally using Friedman ANOVA to characterize the changes in pain intensity within the subjects over time. To compare the efficacy of intervention vs. control, AUCs summarizing the full pain profiles will be compared using the Wilcoxon rank test. Categorical variables will be compared within subjects using Mcnemar's test, and the difference in the incidence of each categorical variable between the intervention and control group will be assessed using chi-square or Fischer exact test as appropriate. Time to recovery will be analyzed using COX regression survival analysis. The hazard ratio (HR) and the corresponding 95% confidence interval (CI) will be reported. Kaplan-Meier curves will be provided to visualize the difference between intervention and control for different time to events. To compare the safety between the two study groups, we will analyze the difference in the incidence of the individual side effects, as well as the composite endpoints of all complications of opioid treatment. Mixed effects regression models will be implemented to account for the complex levels of variability in this study, including within-subject, between-subject, and between-center variability. Fixed effects will include the treatment group, time point, and their interaction. Random effects will consist of hospital and patient-level intercepts. Linear mixed effects models will be utilized for analyzing continuous dependent variables (e.g., AUC for pain profiles), and logistic mixed effects models will be used for analyzing categorical outcomes (e.g., patient satisfaction). Subgroup analysis based on the patient's age will be considered to compare the outcomes between adults and elderly patients (if applicable). Missing data will be removed and will not be included in the analysis for robust estimations. All analyses will be conducted using the intention-to-treat (ITT) principle. A two-sided p-value < 0.05 will be considered statistically significant for all analyses. All statistical analyses will be performed using R version 4.3.1 (R Foundation for Statistical Computing, Vienna, Austria).