OBSTETRICS Postpartum Pain Relief: A Randomized Comparison of Self-Administered Medication and Administration Nathalie East, MD, FRCSC, Johanne Dubé, MD, FRPSC, Élaine Perreault, RN, MSc Department of Obstetrics and Gynaecology, Centre hospitalier de l Université de Montréal (CHUM), Hôpital St-Luc, Montréal, Québec Abstract Objective: To compare pain relief in postpartum women receiving analgesia administered by nurses with the relief achieved by use of self-administered medication (). Method: This randomized trial included women in spontaneous or induced labour or admitted for elective (CS). Women were stratified according to the mode of delivery. In the standard group, nurses administered the medications. In the group, women kept the medications at the bedside and recorded each dose. The women were asked to record their pain level with a validated scale before and one hour after the administration of analgesia. A global pain score was calculated using the mean of these scores, and the satisfaction of patients and nurses with the process was recorded. Results: We recruited 345 women for the study. Eleven women (3.15 %) were lost to follow-up. We analyzed data from 197 women in the vaginal delivery arm and 133 women in the CS arm. There was no difference between the global pain scores before and one hour after the administration of analgesia in each group, independent of the mode of delivery. Women who used and had a vaginal delivery were significantly more likely to have used no medication (P =0.02) or to have used acetaminophen (P = 0.008), and fewer of these women took naproxen (P = 0.05). No significant difference was seen in women who had CS. Narcotic use was similar in each group. Women who used were more likely to indicate that they would choose this method again and recommend it to others. The level of satisfaction with expressed by nurses did not change after the study. Conclusion: Postpartum pain relief was similar for women who had standard administration of medication by nurses and those who had. More women using used either no medication or acetaminophen only, and more women using were highly satisfied with their method of pain relief. Use of self-administered medication should be considered for every postpartum unit. Key Words: Postpartum, pain, self-medication, self-administered medication, patient satisfaction Competing Interests: None declared. Received on October 17, 2006 Accepted on January 18, 2007 Résumé Objectif : Comparer l analgésie chez les femmes en post-partum qui bénéficient d analgésiques administrés par des infirmières à celle que permet l autoadministration de la médication (AAM). Méthode : Cet essai randomisé comprenait des femmes en travail spontané ou déclenché, ou encore hospitalisées en vue d une césarienne de convenance. Les femmes ont été stratifiées en fonction du mode d accouchement. Dans le groupe «standard», les infirmières se chargeaient d administrer la médication. Dans le groupe «AAM», les femmes disposaient de la médication à leur chevet et consignaient la prise de chaque dose. Nous avons demandé aux femmes d évaluer, au moyen d une échelle validée, leur niveau de douleur avant l administration de l analgésie et une heure par la suite. Un score de douleur global a été calculé en utilisant la moyenne de ces scores et la satisfaction des patientes et des infirmières envers le processus a été consignée. Résultats : Nous avons recruté 345 femmes aux fins de l étude. Onze d entre elles (3,15 %) n ont pu faire l objet d un suivi. Nous avons analysé les données de 197 femmes du groupe «accouchement vaginal» et de 133 femmes du groupe «césarienne». Nous n avons constaté aucune différence entre les groupes, en matière de scores de douleur globaux avant l administration de l analgésie et une heure par la suite, et ce, peu importe le mode d accouchement. Les femmes ayant utilisé l AAM et ayant connu un accouchement vaginal étaient considérablement plus susceptibles de n avoir utilisé aucune médication (P = 0,02) ou d avoir utilisé de l acétaminophène (P = 0,008); de plus, un nombre moindre de ces femmes ont eu recours au naproxène (P = 0,05). Aucune différence significative n a été constatée chez les femmes ayant subi une césarienne. Le recours aux narcotiques était semblable dans chacun des groupes. Les femmes ayant utilisé l AAM étaient plus susceptibles d indiquer qu elles choisiraient cette méthode à nouveau et la recommanderaient à d autres. Le niveau de satisfaction envers l AAM exprimé par les infirmières n a pas changé à la suite de l étude. Conclusion : L analgésie post-partum s est avérée semblable chez les femmes ayant bénéficié d une administration standard de la médication par des infirmières et chez celles qui ont eu recours à l AAM. Un plus grand nombre de femmes ayant fait appel à l AAM n ont utilisé aucune médication ou que de l acétaminophène et ont exprimé une grande satisfaction envers le mode d analgésie leur ayant été attribué. Le recours à l autoadministration de la médication devrait être envisagé dans toutes les unités de soins post-partum. J Obstet Gynaecol Can 2007;29(12):975 981 DECEMBER JOGC DÉCEMBRE 2007 975
INTRODUCTION In our experience, many pregnant women request that care during pregnancy, delivery, and the postpartum period be as natural as possible, with minimal use of medication. Allowing women to control their own use of medication by keeping medication at the bedside and taking it when they feel it is necessary (after instruction by nursing caregivers) has been described in obstetrics since 1959. 1 Such use of self-administered medication in women during labour is associated with a high level of satisfaction 2-4 and less use of narcotics. 5,6 However, the implementation of is still limited. Use of gives women more control of postpartum care, and this is likely to have a positive effect on their experience of childbirth. We conducted a randomized trial to compare the experience of women using with the experience of women using standard administration of analgesics. The primary outcome measure was pain relief using the two methods. The secondary outcomes were the rate of use of different medications and the satisfaction of nurses and patients with each form of administration. METHODS Women with a live singleton or twin pregnancy admitted to the CHUM Hôpital St-Luc Birthing Centre (a secondary level centre with 2500 deliveries per year) in labour, for induction of labour, or for elective CS between January 2005 and June 2006 were invited by the attending resident to participate in a study of postpartum analgesia. Those who agreed to participate gave written informed consent. The exclusion criteria were (1) age younger than 18 years, (2) antecedent or current drug abuse, (3) allergy to one of the medications, (4) inability to read and understand French, and (5) inability to give informed consent. After delivery, participants in the study were randomized to treatment group and stratified according to the mode of delivery (vaginal or CS). Randomization was carried out using http://www.randomization.com to provide blocks of 20 patients with an equal opportunity to receive either standard analgesia administered by a nurse or self-administered analgesic medication. Sequentially numbered, opaque, sealed envelopes containing group assignment, prescriptions, and patient information were opened after the women delivered. Since the women and nursing staff knew CS ABBREVIATIONS self-administered medication who had the medications at their bedside, blinding was impossible. The primary outcome measure was pain relief in each group. For this purpose, a 10-point validated scale (0 = no pain, 10 = worst pain) 7 was used in each group. Pain was assessed before and one hour after medication was taken. Secondary outcomes were evaluation of each medication used, the number of tablets taken, the frequency with which tablets were taken, how many women took each medication, and assessment of satisfaction on the part of patients and nurses. Research Design If assigned to the standard administration group, women received their medication upon request from a nurse who recorded the information on a sheet in the patient s file. If they were in the group, the women received a booklet giving information about the medications, secondary effects, and administration schedule. The nurses described the procedure, and the patients recorded each dose on a sheet at their bedside. The nurses were asked to check the patient s comprehension and level of pain relief once on each nursing shift (i.e., every 8 hours), and to ensure that no analgesic medication was unaccounted for. If pain relief was not optimal, the women were prescribed other analgesics, but their outcome data were analyzed in the group to which they were randomized. Self-administration packages were prepared by the pharmacy and kept in the postpartum unit to allow use outside pharmacy hours. Since women who delivered by CS were expected to experience more pain, the packages for these women contained more medication. Each package contained acetaminophen 350 mg (8 or 12 tablets), acetaminophen 350 mg with codeine 30 mg (6 or 16 tablets), naproxen 250 mg (8 or 14 tablets), and sodium docusate 100 mg (8 or 10 tablets). The women were instructed to keep their medications safe and not to share them. Women who used all the medication in a package would be provided with a new one. All elective s were performed under regional anaesthesia using a standard protocol. Regional or general anaesthesia was used for s preformed during labour. Every woman undergoing CS received rectal indomethacin (100 mg) and subcutaneous hydromorphone (1 2 mg at 3 4 hour intervals as needed) for the first 24 hours after delivery; subsequently the or standard analgesic administration began. Women in each group were asked to record their pain levels with a validated visual numerical analogue scale. Pain was assessed immediately before and one hour after the medications were taken. In the standard administration group, the 976 DECEMBER JOGC DÉCEMBRE 2007
Postpartum Pain Relief: A Randomized Comparison of Self-Administered Medication and Administration Flow diagram of subject progress in the study nurse recorded the patient s pain level, and in the group, the patient recorded it herself. The number of tablets taken was also recorded. To compare pain relief in the two groups, a global pain score was calculated for each patient from the mean of all the recorded pain scores before and one hour after administration of medication. At the end of their hospital stay, the study participants were asked to complete a 12-item questionnaire addressing pain relief and satisfaction on a 3-point scale (agree, don t know, disagree). Nurses were asked to complete a 12-item questionnaire with a similar response scale before and at completion of the study to evaluate their perceptions about standard administration of medication and. These questionnaires were completed anonymously. Both questionnaires were designed for the study, but the questions were similar to those used in prior publications on. 2,3 The psychometric properties of these questionnaires were not evaluated. Sample size analysis, performed before the study, indicated that with 80% power and an alpha error of 0.05%, 175 patients would be required in each group to show a difference of 30% in global pain score. The chi-square test was used for discontinuous data, and Student t test was used when appropriate. SAS (version 8.2, SAS Institute Inc., Cary, NC) was used for the statistical analyses. Patient files and pain records were reviewed by the principal author, who was aware of the randomization. Statistics were analyzed by an independent statistician, who was blinded to the group allocation. Ethics approval for the study was given by the Scientific Evaluation and Ethics Committees of CHUM Hôpital St-Luc. RESULTS Pain Three hundred forty-five women were randomized for the study; 200 had a vaginal delivery, and 145 had a Caesarean section. Fifteen patients were lost to follow-up or were excluded. Complete data were available for 161 women in the standard administration group (96 of whom had a vaginal delivery) and for 169 women in the group (101 of whom had a vaginal delivery) (Figure 1). Three women who had a CS did not have sufficient pain relief after receiving study medications and subsequently received other analgesics (2 in the standard group and 1 in ). Seven women had CS performed under general anaesthesia (4 in the standard group, 3 in ). The demographic characteristics of the population studied are shown in Table 1. The study population was multiethnic, with no difference in ethnicity between groups. Data on the pain score at the time of medication administration were available for 171 women (87%) who had a vaginal delivery (88% standard, 86% ) and for 99 women (74%) who had CS (75% standard, 74% ). Data at one hour after the medication were available for 113 women who had a vaginal delivery (57%) and 49 women who had CS (37%), No differences were seen in the pain score before and one hour after medication according to mode of delivery or mode of administration (Table 2). The mean number of tablets used and the mean number of tablets taken for each group are shown in Table 3. In women who had a vaginal delivery, significantly more acetaminophen and less naproxen was used by women in the group. In women who had CS, there was no difference in use of analgesia between the two methods of administration. The DECEMBER JOGC DÉCEMBRE 2007 977
Table 1. Demographic characteristics of study participants (n = 96) (n = 101) P (n = 65) (n = 68) P Maternal age in years (mean SD) 29.5 5.04 29.7 5.06 0.83 31.02 6.27 31.28 6.47 0.81 Nulliparous (%) 49 (51.04) 55 (54.46) 0.63 27 (41.54) 37 (54.41) 0.13 Gestational age in weeks (mean SD) 38.8 4.25 39.3 1.25 0.24 38.3 5.01 38.8 1.37 0.49 Birthweight in grams (SD) 3419 457 3338 474 0.23 3344 839 3361 840 0.91 Maternal hemoglobin 10 g (%) 17 (17.71) 17 (16.83) 0.87 30 (46.15) 31 (45.59) 0.95 Use of regional anaesthesia (%) 73 (78.49) 71 (73.96) 0.14 Duration of labour in hours (mean SD) 10.65 8.31 8.93 5.01 0.08 Spontaneous delivery (%) 81 (84.4) 89 (88.1) 0.45 3rd and 4th degree laceration 8 (8.3) 2 (2.0 ) 0.13 Student t test was used for continous data and chi-square was used for discontinous data. Table 2. Global pain scores before and after medication by mode of delivery (n = 197) (n = 133) (n = 84) (n = 87) P (n = 57) (n = 56) P Global pain score before taking medication Global pain score one hour after medication 4.82 (1.45) 4.40 (1.74) 0.09 5.24 (1.09) 4.91 (1.58) 0.19 (n = 49) (n = 50) P (n = 24) (n = 25) P 2.55 (1.95) 2.31 (2.22) 0.51 2.68 (1.21) 3.09 (1.82) 0.36 All data are shown as mean (standard deviation). Data were compared using Student t test. number of women taking medication in each group is shown in Table 4. Satisfaction Questionnaires were completed by 149 women who had a vaginal delivery (76%) and by 91 women who had CS (68%). Questions and responses are shown in Table 5. Among those who delivered vaginally, fewer women using than using standard administration felt that pain interfered with breastfeeding and learning. In both vaginal delivery and CS groups, women who had preferred to keep their medication at the bedside, would choose this method again, would recommend it to a friend, or stated that they felt in control of their pain relief. Fifty-three nurses completed the questionnaire before the study and 50 after the study. No significant differences were seen in the nurses responses before and after the study. The questions and the nurses responses are shown in Table 6. DISCUSSION In this randomized trial comparing postpartum pain relief in groups of women receiving analgesics administered in one of two ways, global pain scores calculated from a validated scale were similar in each group, indicating that pain relief is similar with standard administration and self-administration of medication, regardless of the mode of delivery. Significantly fewer women who used than women who used standard analgesic administration after vaginal delivery stated that their pain interfered with breastfeeding and learning. Nevertheless, quantitative and qualitative assessment of pain may be different for some women, and even though pain scores are similar, some clinical differences may still exist. 978 DECEMBER JOGC DÉCEMBRE 2007
Postpartum Pain Relief: A Randomized Comparison of Self-Administered Medication and Administration Table 3. Number of tablets and doses taken in each group (n = 96) (n = 101) P (n = 65) (n = 68) P Number of tablets taken Total 7.45 (5.17) 7.01 (4.92) 0.54 11.56 (7.22) 11.12 (7.59) 0.75 Acetaminophen 0.85 (1.72) 1.68 (3.01) 0.009 0.66 (1.79) 1.15 (2.51) 0.20 Acetaminophen with codeine 3.10 (3.95) 2.43 (2.93) 0.18 6.68 (5.84) 6.20 (5.64) 0.64 Naproxen 3.5 (2.73) 2.90 (2.27) 0.07 4.25 (2.59) 3.82 (2.90) 0.38 Number of doses Total 5.88 (3.83) 6.13 (4.30) 0.66 7.80 (5.13) 7.91 (5.89) 0.91 Acetaminophen 0.42 (0.86) 1.17 (1.63) 0.0001 0.28 (0.71) 0.59 (1.42) 0.11 Acetaminophen with codeine 1.84 (2.20) 2 (2.80) 0.63 3.72 (3.37) 4.01 (3.63) 0.62 Naproxen 3.61 (2.57) 2.96 (2.35) 0.06 3.81 (2.76) 3.31 (2.85) 0.30 All data are shown as mean (standard deviation). Data were compared using Student t test. Table 4. Number of women who used each medication Number of patients (n = 96) (n = 101) P (n = 65) (n = 68) P No medication 3 (3.13) 12 (11.88) 0.02 9 (13.34) 14 (20.00) 0.29 Acetaminophen 26 (27.08) 51 (50.50) 0.001 12 (17.65) 20 (28.57) 0.13 Acetaminophen with 59 (61.46) 58 (57.43) 0.56 53 (77.94) 51 (72.86) 0.49 codeine Naproxen 88 (91.67) 78 (77.23) 0.005 55 (80.88) 50 (71.43) 0.19 Data were compared using chi-square test. We were interested to find that women delivering by CS reported a pain level only slightly higher than the level reported by women who delivered vaginally. Although our study was not designed to compare the level of pain according to type of delivery, we expected that women who underwent CS would have a higher global pain score. That they did not might be explained by the fact that the pain score was not recorded immediately after CS but 24 hours later. Our finding that women who delivered vaginally and were randomized to used more acetaminophen and less naproxen may not have been clinically significant, since the total number of tablets was similar. Women in the group might have used more acetaminophen simply because they were familiar with it and had used it previously at home. They might also have wanted to avoid the drowsiness caused by codeine. Before the study, the nurses distributed naproxen regularly to patients, and they might have continued to suggest its use more freely during the study. Significantly more women in the group who delivered vaginally chose to use no medication at all, most likely because they were empowered to make this choice. Before the study, the nursing staff expressed concern about the possibility of narcotic abuse if the women kept their medications at the bedside. During the study, no loss of tablets and no abuse of acetaminophen with codeine were reported in the group. The number of women who elected to take acetaminophen with codeine, the number of tablets taken, and the number of doses were similar in both study groups. In a retrospective study of postpartum pain relief, Greene et al. 5 reported a significant decrease in narcotic use (9.23 vs. 2.01 tablets, P 0.001) and in the proportion of patients who used any narcotics (82.5% vs. 40%) in regimens. Macartney and Whyte 6 reported a reduction of 85% in the use of acetaminophen with codeine 30 mg DECEMBER JOGC DÉCEMBRE 2007 979
Table 5. Patient satisfaction questionnaire results Question (%) (%) P (%) (%) P I like the way I received/took my medication. 57/67 (85.07) 74/80 (92.5) 0.29 33/42 (78.57) 45/49 (91.84) 0.19 I received/took my medication when I needed it. 61/67 (76.25) 75/80 (93.75) 0.48 35/42 (83.33) 43/49 (87.76) 0.61 I feel ready to take my medication alone at home. 65/67 (97.05) 77/80 (96.25) 0.55 40/41 (97.56) 49/49 (100) 0.27 My pain relief was appropriate during my stay. 59/68 (86.76) 73/81 (90.12) 0.61 37/42 (88.10) 45/48 (93.75) 0.46 My pain interfered with my learning. 9/68 (13.24) 6/81 (7.40) 0.04 7/42 (16.67) 5/49 (10.20) 0.37 My pain interfered with breastfeeding. 11/67 (16.42) 3/79 (3.80) 0.03 7/39 (17.95) 4/49 (8.16) 0.21 I am satisfied with the way I received my medication. 58/67 (86.57) 73/80 (91.25) 0.64 32/42 (76.19) 44/49 (89.80) 0.13 I prefer when the medications are distributed by the 28/66 (42.42) 5/80 (6.25) 0.001 16/41 (39.02) 9/49 (18.37) 0.008 nurses. I prefer when the medications are kept by the 26/63 (41.27) 61/73 (83.56) 0.001 15/40 (37.50) 35/48 (72.92) 0.003 bedside. I would choose the same method again. 40/63 (63.63) 70/75 (93.33) 0.001 23/40 (57.50) 42/48 (87.50) 0.005 I would recommend this method to a friend. 42/64 (65.63) 68/75 (90.660 0.001 24/40 (60.00) 38/47 (80.35) 0.03 I felt in control of my pain relief. 48/64 (75.00) 72/76 (94.74) 0.003 26/40 (65.00) 46/48 (95.83) 0.001 Numerators are the number of women who circled the answer agree, and denominators are the number of women who answered the question. Only the agree results are shown. tablets after implementation of a self-medication program for postpartum patients. In these two studies, the self-administration package did not include narcotics, and this could have contributed to the magnitude of the reduction in use of narcotics. Patients might have been reluctant to ask for medication that was stronger than usual. In another retrospective postpartum study in which the package included acetaminophen with codeine 30 mg tablets, Anderson and Poole 8 reported a significant increase in use (from a mean of 2 tablets to 4 tablets, P = 0.01). Although the protocol used in this study was similar to that used in our study, we did not find a significant increase in use by self-medicating women. The fact that no narcotic abuse or increase in consumption occurred in our population should reassure caregivers involved in similar programs. The questionnaire results showed that women in both groups were satisfied with the medication method to which they were assigned. Women in the group clearly preferred to have their medication at their bedside. The final questions were designed to address satisfaction more precisely, and we found that more women allocated to would choose it again and would recommend it to a friend. They also felt more in control of their pain relief, consistent with the empowerment that gives to women. We anticipated that this empowerment would have a stronger impact on the pain level than it did. The fact that we found no differences in the pain score or medication use for the CS group might be because our sample size was smaller in the CS arm than in the vaginal delivery arm. Nevertheless, the women who had CS and were randomized to were as satisfied as those in the vaginal delivery group with the degree of pain relief. This demonstrates that is readily accepted by all patients. We anticipated that after experiencing, nursing staff would be enthusiastic about it, but no significant difference was found between nurses opinions before and after the study. In fact, fewer nurses felt that patients were informed and knew how to take their medications after the project than before it, although the difference was not statistically significant. At the end of the project, more nurses stated that they preferred standard administration, and fewer nurses stated that they preferred. During the project, nurses had to explain the booklet to patients, confirm their understanding of it, and become used to a new care model; these additional requirements might have contributed to their lower opinion of. The findings of the study, and especially the high patient satisfaction level, should serve to reassure them. The study has some limitations, particularly concerning its internal validity. Eleven patients were randomized but not properly identified, and they could not be tracked. Pain recording was not always carried out by the nurses and the 980 DECEMBER JOGC DÉCEMBRE 2007
Postpartum Pain Relief: A Randomized Comparison of Self-Administered Medication and Administration Table 6. Results of the nurses questionnaire Question Before (%) (n = 53) After (%) (n = 50) P Patients are well informed about their medication. 33/53 (62.66) 26/50 (52.00) 0.54 Patients understand when they have to take their medication. 28/52 (53.85) 23/50 (46.00) 0.73 Patients received their medication rapidly. 38/53 (71.70) 42/50 (84.00) 0.29 Patient s pain is well relieved. 38/53 (71.70) 41/50 (82.00) 0.40 I prefer the standard administration of medication. 15/51 (29.41) 20/50 (40.00) 0.44 administration is a good method of medication. 22/52 (42.31) 25/50 (50.00) 0.70 administration gives good pain relief. 36/52 (69.23) 33/48 (68.75) 0.37 I prefer. 26/47 (55.32) 24/48 (50.00) 0.43 is a good method of medication. 30/47 (63.83) 28/48 (58.33) 0.42 gives good pain relief. 35/47 (74.47) 31/48 (64.58) 0.58 gives better pain relief than standard administration. 21/47 (44.68) 22/48 (45.83) 0.68 Numerators are the number of nurses who circled the agree answer, and denominators are the number of nurses who answered the question. Only the agree results are shown. patients, although they had written instructions on how to record pain. We did not record how many women declined to participate in the study. Our trial was generally well accepted, but it would have been instructive to compare the characteristics of women who declined to participate with those who agreed. CONCLUSION The results of our study indicate that postpartum pain relief in a program was similar to that recorded by women receiving standard administration of analgesics. Women who delivered vaginally and who self-administered medication used more acetaminophen and less naproxen than women receiving standard administration, and a larger proportion did not medicate at all. The use of narcotics was similar in both groups of women. Patients in the group were highly satisfied with this method of administration. We believe that all postpartum units should consider the implementation of in order to increase patient satisfaction and empowerment. ACKNOWLEDGEMENTS The editorial assistance of M Ovid Da Silva and the statistical work of Ms Bin Wei are acknowledged. REFERENCES 1. Parnell MA. Medicines at the bedside. Am J Nurs 1959;59:1417 8. 2. Jankowski H, Wells SM. Self-administered medications for obstetric patients. MCN Am J Matern Child Nurs 1987;12(3):199 203. 3. Hoffmann RP, Kerchner J, Osburn RC, St Louis P, Stewart H. A self-medication program for obstetric patients. Hosp Pharm 1978;13(3):127,130 2,134. 4. Nynick A. Instituting a postpartum self-medication program. MCN Am J Matern Child Nurs 1981;6(6):422 4. 5. Greene JF, Kuiper O, Morosky M, Wightman S, Curry SL. A postpartum self-medication program: effect on narcotic use. J Womens Health Gend Based Med 1999;8(8):1073 6. 6. Macartney S, Whyte M. New moms and self medication. Can Nurse 1995;91(9):47 8. 7. Wong DL, Baker CM. Pain in children: comparison of assessment scales. Pediatric Nursing 1988; 14 (1):9 17. 8. Anderson K, Poole C. Self-administered medication on a postpartum unit. Am J Nurs 1983;83(8):1178 80. DECEMBER JOGC DÉCEMBRE 2007 981