Hospitalized patients’ pain experience before and after the introduction of a nurse-based pain management programme: a separate sample pre and post study

A quasi-experimental design with a separate sample pretest-posttest approach was conducted at Jimma University Medical Center (JUMC) from 1 September 2016, to 15 July 2017.

A convenience sample of 845 patients (Survey 1: N = 282; Survey 2: N = 283; Survey 3: N = 280) was invited to participate in this study. All patients who were admitted to the four inpatient units at the hospital (medical, surgical, maternity, and gynecology wards) were included if they filled the inclusion criteria. Patients had to have been hospitalized for at least 24 h, age ≥ 18 years, and have no known hearing impairment. Initial contact with patients was made through the ward’s head nurse or shifts leader. Participation in the study was voluntary. None of the patients approached declined participation.

A nurse-based pain management programme is an intervention comprised of two components: education (to enhance nurses’ knowledge of and attitude towards pain) and organizational elements (to ensure the systematic monitoring of patients’ pain), with the goal of improving pain treatment. In the educational component, we provided two days of intensive in-service pain management training (16 h of face-to-face training), take-home reading assignments (self-learning), and refresher training four weeks later (8 h) for all nurses in the units. The education programme was arranged in groups and completed between 1 October and 15 November 2016. Each group was comprised of 30–40 nurses. The education sessions were delivered as per the pain management protocol developed by the research team, which was based on Ethiopia’s Federal Ministry of Health (FMOH) pain management guidelines [11] the World Health Organization’s (WHO) guidelines for pain management [26, 27]. Following the education programme, we introduced the second component, a rounding programme to educate staff nurses and nurse leaders in patient goal-oriented pain management [28]. The rounding programme consisted of an engagement orientation on how to organize and conduct rounding. The rounding programme lasted one day (8 h) for all staff nurses and a half day (four hours) for nurse leaders and supervisors. The content of the orientation included regular pain assessment using the numerical rating scale (NRS), charting in rounding logs when it was necessary to consult the physician, scripted dialogue with the patients, and how to assess patients using the four Ps: presence, pain, position, and personal needs. Rounding was structured in such a way that nursing directors, head nurses, and staff nurses proactively made regular and consistent visits to patients and performed scheduled tasks focused on pain management. When nurses visited patients during rounds, they introduced themselves and read the following script in either the Afaan Oromo or Amharic language (according to the patient’s language preference): “We are going to do everything we can to help keep your pain under control. Your pain management is our number-one priority. Given your (condition, history, diagnosis, status), we may not be able to keep your pain level at zero. However, we will work very hard with you to keep you as comfortable as possible” [4]. Staff nurses made subsequent visits every 2 h during the day (8,00 am–8,00 pm) and every 4 h at night (10,00 pm–6,00 am). During each visit, nurses assured patients of their availability and informed patients when they would return, asked patients to rate their pain levels, and recorded it in the pain log. If necessary, they repositioned the patient and checked for personal needs (toileting, getting out of bed, water).

More specifically, staff nurses systematically assessed every patient admitted to the four units up to ten times (every 2 h during the day and every 4 h during the night) in a 24-h period. Unless the patient was unconscious, sleeping or the bed was empty, the pain level was self-rated using the NRS and recorded in the pain log by the nurse. After each pain assessment, the nurse decided if the patient required a change in pain treatment regimen in collaboration with the treating physician, using the WHO pain ladder framework. Based on the collaborative decision, the nurse administered adjusted pain medication by the clock. “By the clock” means that the patient would be given analgesics regularly at a fixed interval of time-based on the known pharmacokinetics of the drug in use and that the next dose of analgesics would be adjusted before the effect of the previous dose had fully worn off. The rounding log was kept easily accessible for the healthcare provider’s review. In addition, nurses participated in the multidisciplinary team rounds and shared patient pain information.

Leadership rounding was performed daily by head nurses or clinical leaders (team leaders) and weekly by nursing directors. The leaders’ role was to motivate, facilitate, and provide positive feedback to the nurses. In addition, the head nurse-led weekly staff nurse discussions, and the nursing director led monthly discussions for head nurses and supervisors. Compliance with the rounding protocol was monitored twice weekly by nurse supervisors, using a review of the rounding log and discussion minutes.

Pain intensity and interference are regarded as reliable parameters to measure patients’ experiences of pain [29]. To measure the patient pain experience, we used a tool consisting of 18 items adapted from the Brief Pain Inventory (BPI) [30] and the American Pain Society Pain Outcome Questionnaire-Revised (APS-POQ-R) [31]. Items that were used to measure pain prevalence in the last 24 h (one item), pain treatment information (two items), and pain intensity (four items) were adopted from the BPI. Pain intensity/severity was measured on the 11-points NRS (from 0 = no pain to 10 = worst possible pain) with four scores: for pain present at the time of interview (“right now”), pain at its worst during the last 24 h, pain at its least over the past 24 h, and pain on average over the past 24 h. A pain severity/intensity level less than 4 is regarded as mild pain, greater than or equal to 4 and less than or equal to 6 is moderate, and greater than or equal to 7 is severe pain on the 0 to 10 NRS [1]. Eleven items used to measure pain interference (six items for physical functions and four items for emotional functions) were adopted from the APS-POQ-R. Patients were asked to rate their pain’s interference with physical functions such as activity in bed (sitting up, turning in bed), activity out of bed (walking, standing, squatting, use of wheelchair, dressing, etc.), deep breathing and coughing exercise (postoperative patients), sleeping (falling asleep, staying asleep), and relationships with others as well as interference with emotional function, in this case mood disturbance (anxious, depressed, frightened, feeling helpless). All functions were measured on an 11-point rating scale (0 = no interference, 10 = complete interference). Scores less than 3 indicate mild, greater than or equal to 3 and less than or equal to 4 indicate moderate, and greater than 4 indicate severe interference [1].

The tool was initially translated by healthcare professionals to Afaan Oromo and Amharic; linguistics and non-health care professionals then retranslated back to English. Then, all the translators came together to discuss the translated items. To check how each item was understood, a cognitive interview was conducted with five people of varying backgrounds [32]. Finally, the tool was tested on 35 patients from various units to clarify words and the sequence of the items. We have also collected information on admission unit (the type of unit), sociodemographic characteristics (age, sex, address), and socioeconomic variables (educational level, occupation, monthly income). The principal investigator collected the completed questionnaires from the data collectors daily and stored them in locked cabinets.

Data was first collected at baseline (Survey 1), again six weeks after the educational programme (Survey 2), followed by a third survey immediately after four months of rounding (Survey 3). The data were collected by trained nurses through a structured face-to-face interview that lasted approximately 40–45 min.

All surveys were assessed for missing or incomplete data before being analyzed (survey 1: 26, survey 2: 24, survey 3: 13). Data were analyzed using Statistical Package for the Social Sciences (SPSS) version 20.1 (IBM SPSS Statistics for Windows, Armonk, NY). Descriptive statistics (i.e., mean, standard deviations, range, frequency) were calculated for patient characteristics (age, income), pain severity, and interference response items. Reduction in the sample means score was calculated by subtracting the mean value of survey 3 from the mean value of survey 1, dividing it by the survey 1 value, and multiplying by 100. Differences between the mean pain intensity and interference scores at baseline (Survey 1), six weeks after the in-service educational program (Survey 2), and immediately after four months of rounding (Survey 3) were analyzed using a one-way Analysis of Variance (ANOVA) with a post-hoc Bonferroni test. The significant differences between the surveys were declared at p < 0.05.

Of the 845 eligible patients, 782 patients’ complete responses from (Survey 1: N = 256; Survey 2: N = 259; Survey 3: N = 267) were analyzed (Fig. 1). There were no differences in the mean ages of the respondents (Survey 1: mean age = 38.1 (SD ± 16.2); Survey 2: mean age = 37.4 (SD ± 15.2); Survey 3: mean age = 37.9 (SD ± 5.2)). Though the percentage of missing values for income was high (30.4% for Survey 1, 42.8% for Survey 2, and 60% for Survey 3), the median monthly income in Ethiopian currency was 1000 birr for all surveys: the interquartile range was 1054 birr for Survey 1, 1500 birr for Survey 2, and 1500 birr for Survey 3. When converted to US currency, 50% of respondents earned just over one US dollar per day (1 US dollar ≈ 27.91 Ethiopian birr). There were no statistically significant differences in sociocultural characteristics between the three samples, and there were no reports of new disease epidemics in the area. As shown in Table 1, except for the unit of admission, there was no difference in the distribution of the sample characteristics by survey period.

Most patients received anti-pain medication intramuscularly and/or intravenously. In Survey 2, the proportion of patients treated with a pharmacological agent increased by 14.1% compared with Survey 1, by 4.8% in Survey 3 compared with Survey 2, and by 19.7% in Survey 3 compared with Survey 1. Prayers, massage, and cold or hot application were commonly reported nonpharmacological pain therapies provided by patient attendants (Table 2).

Table 3 shows the samples’ mean pain intensity in the first, second, and third surveys. The results of all three surveys show that patients generally had moderate to severe pain. However, the mean pain intensity levels were generally reduced across the survey period. This reduction was statistically significant between the second and first survey as well as between the third and first survey for the worst pain, least pain, and pain “right now”. Reduction in the samples’ mean scores for pain intensity at its average during the last 24 h was statistically significant between all three surveys. In the third survey, the sample means’ pain intensity was reduced by 27.6% at for pain at its worst, 23.8% for pain at its least, 25.5% for pain at its average over the last 24 h, and current pain by 29.3% compared with Survey 1. Even though the proportion of patients who reported pain in the last 24 h generally decreased after the intervention, the proportion of patients who experienced pain in the last 24 h in the second survey was slightly higher (94.5%) than in the first survey (93%). However, immediately after four months of rounding (in the third survey), the proportion was reduced to 87.3%.

The results of all three surveys show that patients generally had moderate to severe pain when asked for “average pain in the last 24 hours” and “pain right now”. However, as indicated in Table 4, the proportion of patients with severe pain was generally reduced across the survey period.

In all three surveys, the score for the mean pain interference scales indicates moderate to severe interference with both physical and emotional functions. As shown in Table 4, only minor differences were observed in the level of patient-reported pain interference between Survey 1 and Survey 2. Apart from activities out of bed, a statistically significant reduction in pain interference with physical functions was observed between both the first and third surveys and the second and third surveys. On the other hand, the mean level interference with relationship and negative feelings (anxious, depressed, frightened, and helpless) significantly decreased in Survey 2 and Survey 3 compared with Survey 1. However, the reduction in feeling helpless occurred between the third and second surveys (Table 5).

The current findings indicate the potential benefits of a nurse-based pain management programme (in-service educational programme and rounding) for hospitalized patients. However, several factors may limit the findings of the current study. One limitation could be attributed to the study design, the use of a non-randomized design without a control group. Initially, we planned to employ a quasi-experimental design with a control group. This was however complicated by a period of public unrest followed by a declaration of a state of emergency, which made it impossible to travel between the intervention and control sites, and our plans to use a control group had to be abandoned. Another limitation is the possibility that nurses may have gained additional knowledge by interacting with medical professionals, thereby improving their practices, resulting in the possibility of physician-initiated pain treatment rather than nurse-initiated pain treatment through consultation. Given all these factors, a simple pre-post study design with three measurement points on separate samples is inadequate for causal inference. Further randomized, multicenter studies are necessary before attributing a nurse-based pain management programme to changes in patient-reported pain intensity and interference. Even though we have no data on the duration of hospitalization or types of procedures the patients had undergone during their hospital stays, it should be noted that this might have influenced the severity of pain.