Implementation a nursing information system to evaluate clinically significant pain prevalence and pain relief among hospitalized cancer patients

104

Background: A nursing information system (NIS) was established to document patient-reported pain intensity. Among cancer inpatients, clinically significant pain (CSP) is one of the most common complaints during repeated hospitalizations. Although the prevalence is the popular indicator, the CSP prevalence and the magnitude of pain relief are not determined. There is still a lack of information on evaluating the pain intensity difference (PID) and the percentage difference in pain intensity (%PID) which can be determined according to the worst pain intensity (WPI) and last evaluated pain intensity (LPI) before discharge. The PID and %PID could calculate the magnitude of pain relief, but these indicators do not consider the goal of pain relief. As a result, a new measurement indicator to evaluate pain relief outcomes is also very important for pain management. Objectives: In terms of previous study results and clinical experience, this study divided into two parts. The aim of the first part was to characterize the trends in CSP prevalence and evaluate the PID and %PID in cancer inpatients with moderate or severe pain during each hospitalization from 2011-2013. Meanwhile, we examine the differences in CSP prevalence and pain intensity across repeated hospitalizations among cancer inpatients. The aim of the second part was to develop a pain relief index (PRI) for pain relief outcome evaluation and to examine the index validity using the demographic characteristics among cancer inpatients with CSP. Methods: Patient-reported pain intensity was assessed using a numerical rating scale, a faces pain scale or the Face, Legs, Activity, Cry, Consolability Behavioral Tool. Pain intensity greater than 4 points were defined as CSP, and a pain intensity ≥ 7 points was defined as severe pain. For each patient during each hospitalization, we selected only one pain score, which included two types: the WPI and the LPI before discharge. Patient-reported pain intensity scores were recorded in an NIS. Using the NIS, a pain score database containing all the cancer patients hospitalized data from 2011 through 2013 was analyzed. The hospital-based, retrospective study was conducted at an academic medical center. Regarding the CSP prevalence and pain intensity, we examined the differences in the prevalence of the WPI and the LPI of ≥ 4 or ≥ 7 points among cancer inpatients from the 1st to the 18th hospitalizations. Linear mixed models were utilized to determine the significant difference in the WPI and LPI (≥ 4 or ≥ 7 points) at each hospitalization. As for the pain intensity, the PID and %PID were examined using the one-sample Kolmogorov-Smirnov test. Group differences in moderate or severe pain with respect to the number of hospitalizations were evaluated using the Mann-Whitney U test. The second part of the study, a new develop PRI to evaluate outcome data among cancer inpatients and the PRI as the dependent variable to examine the demographic characteristics. The %PID and PRI were examined using the one-sample Kolmogorov-Smirnov test. The group differences of the demographic characteristics for continuous data comparisons were evaluated using the Mann-Whitney U tests. Linear regression was performed to examine the relationships between the demographic characteristics and the PRI. The dependent variable in the regression equation was the PRI, and the independent variables were the demographic characteristics. Spearman’s correlation coefficient analysis was performed to evaluate the criterion-related validity of the PRI. Results: We determined the CSP prevalence and pain intensity difference during each hospitalization, and the number of hospitalizations per patient at this academic medical center ranged from one to 18, and 88,133 pain scores were studied. The prevalence of a WPI ≥ 4 points during the 1st hospitalization was significantly higher (0.240-fold increase) than the 5th hospitalization. For the 2nd, 3rd and 4th hospitalization, there was a significantly higher prevalence of a WPI ≥ 4 points compared to the 5th hospitalization. We also found significant decrease in the prevalence of WPI ≥ 7 points during the 1st to the 4th hospitalizations, LPI ≥ 4 points during the 1st to the 3rd hospitalizations, and LPI ≥ 7 points during the 1st to the 2nd hospitalizations. All the prevalence of four CSP types showed a trend toward a decrease from the 1st to the 18th hospitalizations. There was a robust decrease in the CSP prevalence from the 1st to the 5th hospitalizations, except for an LPI ≥ 7 points. As for the pain intensity, among patients with a WPI of 4 to 6 points, from the 1st to the 18th hospitalizations, the mean PID was 1.52 and the %PID was 29.0%, both of which were significant. There were more significant changes in the PID and %PID associated with a WPI ≥ 7 points. The average PID in the WPI ≥ 7 group (3.09) was significantly higher than the average PID of the 4 ≤ WPI ≤ 6 group (1.52). The average %PID of the WPI ≥ 7 group (38.5%) was also significantly higher than the average %PID of the 4 ≤ WPI ≤ 6 group (29.0%). In addition, the second part of the study, we focused on cancer inpatients with CSP for whom the WPI was ≥ 4 points. PRI values of -62.02% to -72.55% were observed in the WPI ≥ 7 and 4 ≤ WPI ≤ 6 groups. Significant (p < 0.05) effects on PRI values were observed among patients who were >65 years, admitted to the medicine or gynecology and had a >30-day hospital stay. Conclusion: Using the NIS, we effectively collected pain intensity scores. This first hospital-based study that used an electronic database to analyze CSP prevalence, pain intensity such as PID, %PID and new index-PRI among cancer inpatients in Taiwan. The hospitalization data indicated that the prevalence of the four CSP types reduced gradually and the cancer inpatients with moderate and severe pain experienced reductions in pain of approximately 30% and 40%, respectively. Because this is a descriptive and observational study, it is impossible to address the causative factors on the basis. Next step will examine the causative factors that determining the CSP prevalence and CSP reduction among cancer patients. In addition, using an electronic NIS to validate the PRI is an effective and valid measure for evaluating outcome data among cancer inpatients. In the future, further defined the meaningful range of percentage difference in the PRI from various perspectives should be performed.