Improving Ovine Behavioral Pain Diagnosis by Implementing Statistical Weightings Based on Logistic Regression and Random Forest Algorithms.

Recently, the Unesp-Botucatu sheep acute pain scale (USAPS) was created, refined, and psychometrically validated as a tool that offers fast, robust, and simple application. Evidence points to an improvement in pain diagnosis when the importance of the behavioral items of an instrument is statistically weighted; however, this has not yet been investigated in animals. The objective was to investigate whether the implementation of statistical weightings using machine learning algorithms improves the USAPS discriminatory capacity. A behavioral database, previously collected for USAPS validation, of 48 sheep in the perioperative period of laparoscopy was used. A multilevel binomial logistic regression algorithm and a random forest algorithm were used to determine the statistical weights and classify the sheep as to whether they needed analgesia or not. The quality of the classification, estimated by the area under the curve (AUC) and its 95% confidence interval (CI), was compared between the USAPS versions. The USAPS AUCs weighted by multilevel binomial logistic regression (96.59 CI: [95.02-98.15]; p = 0.0004) and random forest algorithms (96.28 CI: [94.17-97.85]; p = 0.0067) were higher than the original USAPS AUC (94.87 CI: [92.94-96.80]). We conclude that the implementation of statistical weights by the two machine learning algorithms improved the USAPS discriminatory ability.

Correction: Validation of the Unesp-Botucatu composite scale to assess acute postoperative abdominal pain in sheep (USAPS).

[This corrects the article DOI: 10.1371/journal.pone.0239622.].

Validation of the Unesp-Botucatu composite scale to assess acute postoperative abdominal pain in sheep (USAPS).

A scale with robust statistical validation is essential to diagnose pain and improve decision making for analgesia. This blind, randomised, prospective and opportunist study aimed to develop an ethogram to evaluate behaviour and validate a scale to assess acute ovine postoperative pain. Elective laparoscopy was performed in 48 healthy sheep, filmed at one preoperative and three postoperative moments, before and after rescue analgesia and 24 hours after. The videos were randomised and assessed twice by four evaluators, with a one-month interval between evaluations. Statistical analysis was performed using R software and differences were considered significant when p <0.05. Based on the multiple association, a unidimensional scale was adopted. The intra- and inter-observer reliability ranged from moderate to very good (intraclass correlation coefficient ≥ 0.53). The scale presented Spearman correlations > 0.80 with the numerical, simple descriptive, and visual analogue scales, and a correlation of 0.48 with the facial expression scale. According to the mixed linear model, the scale was responsive, due to the increase and decrease in pain scores of all items after surgery and analgesic intervention, respectively. All items on the scale demonstrated an acceptable Spearman item-total correlation (0.56-0.76), except for appetite (0.25). The internal consistency was excellent (Cronbach's α = 0.81) and all items presented specificity > 0.72 and sensitivity between 0.61-0.90, except for appetite. According to the Youden index, the cut-off point was ≥ 4 out of 12, with a diagnostic uncertainty zone of 4 to 5. The area under the curve > 0.95 demonstrated the excellent discriminatory capacity of the instrument. In conclusion, the Unesp-Botucatu pain scale in sheep submitted to laparoscopy is valid, reliable, specific, sensitive, with excellent internal consistency, accuracy, discriminatory capacity, and a defined cut-off point.