Risk Factors for Impaired Patient-Reported Satisfaction and Increased Length of Hospital Stay Following Hysterectomy on Benign Indications in Premenopausal Women: a Study From the Swedish National Register for Gynecological Surgery

Fig. 2 presents a flow chart of the study population. Flow chart demonstrating the selection of the study population of women undergoing hysterectomy. Demographic, descriptive and clinical characteristics subdivided according to mode of surgery at conclusion of the operation are demonstrated in Table 1 . The number of missing data is presented in Tables 2 and 3 in connection with each variable and varies from a few per cent to a larger proportion; in the latter case mostly due to the limitation in use over time of the item in the inquiry forms. Five hundred and thirty-six of those who were originally scheduled for a minimally invasive procedure, 282 (4.4%) laparoscopically and 254 (7.1%) vaginally, were intraoperatively converted to abdominal hysterectomy. Table 1  Demographic, descriptive and clinical data of 27 938 women undergoing hysterectomy for benign conditions in relation to mode of surgery (at conclusion of the operation). Yes No Missing data Menstrual disorder Uterine myoma Pain-related* Other indications Missing data Class I/II Class III/IV Missing data Yes No Missing data Get rid of symptoms Relieved None Missing data § Table 2  Occurrence and classification of complications intra- and postoperatively after hysterectomy in relation to surgical mode. None Minor Major Missing None Minor Major Missing None Minor Major Missing None Minor Major Missing § Table 3  Complications intra- and postoperatively after hysterectomy on various occasions until one year postoperatively in relation to dichotomized self-reported perception of medical condition/result of surgery. None Minor Major Missing data # None Minor Major Missing data # None Minor Major Missing data None Minor Major Missing data Prevalence and degree of severity of complications in relation to the occasion of reporting the complication and to modes of hysterectomy are shown in Table 2 . In total, 67.9% (11 132/16 388) reported a completely complication-free hysterectomy within the first year postoperatively. Table 3 shows the self-reported perception of medical condition and opinion of result of surgery one year postoperatively in relation to reported complications and severity intra- and postoperatively. The women with a complication-free hysterectomy reported an improved medical condition and satisfaction with the hysterectomy in 95 – 97% and 93 – 95%, respectively, but the prevalence decreased to 80 and 60%, respectively when major complications had occurred. The associations between complications and patient satisfaction reported one year postoperatively are shown in Table 4 . When adjusted for confounding factors, medical condition after one year was strongly negatively associated with complications of all grades occurring after discharge from hospital. Except for minor complications occurring intraoperatively, the reported satisfaction with hysterectomy one year after surgery was strongly negatively associated with all grades of severity of complications on all occasions. In particular, major complications reported at the one-year inquiry were highly associated with dissatisfaction with the result of surgery. Table 4  Associations between complications and the dichotomized self-reported perception of medical condition and satisfaction on various occasions until one year postoperatively. None Minor Major None Minor Major None Minor Major None Minor Major Table 5 shows various predictive factors for unfavorable outcome of surgery reported one year postoperatively. In the adjusted models, the perceived medical condition was significantly adversely associated with having no preoperative expectation of surgery (aOR 4.85 [95% CI 2.80 – 8.40]), smoking (aOR 1.45 [95% CI 1.14 – 1.84]), having uterine myoma (aOR 2.33 [95% CI 1.82 – 2.98]), or pain-related (aOR 3.29 [95% CI 2.57 – 4.22]) as indications for hysterectomy. Of the mode of hysterectomy only vaginal hysterectomy was significantly positively associated with the medical condition reported at the one-year inquiry (aOR 0.64 [95% CI 0.47 – 0.87]). Furthermore, vaginal and subtotal abdominal hysterectomies were independent predictive factors for higher satisfaction after surgery (aOR 0.61 [95% CI 0.51 – 0.71] and aOR 0.74 [95% CI 0.62 – 0.87], respectively). Pain as indication of surgery was a strong risk factor for experienced dissatisfaction (aOR 2.17 [95% CI 1.85 – 2.55]). Likewise, smoking and having no preoperative expectations of surgery or believing that the only benefit of surgery would be alleviated of symptoms and being without ovaries after the surgery were independently associated with dissatisfaction. Intraoperative conversion of a minimally invasive procedure to laparotomy did not seem to affect the reported medical condition or satisfaction with the surgery one year after the operation. Table 5  Risk factors for unfavorable outcome of the surgery concerning self-reported perception of medical condition and satisfaction one year postoperatively. Results of logistic regression analyses with dichotomized outcome measures. Abdominal total hysterectomy Abdominal subtotal hysterectomy Laparoscopic total hysterectomy Laparoscopic subtotal hysterectomy Vaginal hysterectomy Menstrual disorder Uterine myoma Pain-related Other indications Get rid of symptoms Relieved No expectations No Yes Yes Class III/IV The LOS for the entire study group was mean 2.3 days (1.9 days). Occurrence of complications intraoperatively and postoperatively before discharge from hospital, and degree of complications had statistically significant impact on LOS ( Fig. 3 a and b ). The LOS was predictive for occurrence and degree of complications reported after discharge ( Fig. 3 c and d ). The post hoc tests revealed that there were highly statistically significant differences in LOS (p < 0.001) between all three groups (complication-free, minor and major complications). The adjusted LOS was mean 2.1 days (1.6 days) for those who had no reported complications on any occasion. This was significantly lower than the mean 2.8 days (2.1 days) (p < 0.001) for those who had a complication of any degree on at least one occasion. Length of hospital stay (LOS) in relation to occurrence of reported complications after hysterectomy. a  Intraoperative complications. b  Complications after surgery during hospital stay. LOS was also adjusted for occurrence of intraoperative complications. c  Complications reported to have occurred between discharge and eight weeks postoperatively. LOS was also adjusted for complications intraoperatively and during hospital stay. d  Complications reported to have occurred between the 8-weeks and 1-year inquiry. LOS was also adjusted for complications intraoperatively, during hospital stay and at the 8-weeks inquiry. The majority of women were satisfied with the result of the hysterectomy but there was an association with decreased satisfaction when complications had occurred. Pelvic pain as indication of surgery, no expectations preoperatively, and current smoking were independent risk factors for adverse outcomes in patient satisfaction. Occurrence of complications intra- and postoperatively before discharge from hospital was associated with increasing LOS. Furthermore, LOS was predictive for occurrence and degree of complications reported after discharge from hospital. A strength of this research is the large study population, based on prospectively collected data in the GynOp. More than 75% of all benign hysterectomies performed in Sweden during the study period formed part of the register. Another strength is that exclusion criteria were chosen to ensure that the indication of hysterectomy was most likely of uterine origin and not related to other pelvic symptoms, diseases or conditions that per se could have an impact on perceived satisfaction. However, risk of selection bias and missing data in the register represent weaknesses of the study. The interpretation of the results should be done with caution due to the inborn risk of multiple testing problems in epidemiological studies. Due to the exploratory nature of this study no statistical corrections of the p-values were done. A drawback of the study and a limitation is the lack of an overall comprehensive validation of the GynOp. However, several articles (> 50 in peer review reviewed journals) and more doctoral dissertations emanate from the register 10 which indicate that the register is considered reliable, probably because of the high internal coverage of all gynecological surgery in Sweden. Prevalence and degree of intra- and postoperative complications in relation to mode of surgery were in accordance with results of previous studies 5 ,  11 . Complications, especially major, were strongly associated with less contentment with medical condition and satisfaction with surgery one year postoperatively. Moreover, pain as indication of surgery, having no or low preoperative expectations of surgery, and current smoking were above all significantly associated with adverse outcomes concerning patient-reported satisfaction. Impact of different risk factors associated with adverse outcomes in patient satisfaction following gynecological surgery have rarely been described previously. Two studies have reported that minimally invasive surgery seemed to be advantageous concerning patient satisfaction following hysterectomy 7 ,  9 . Billfeldt et al. who also used data from GynOp, excluded women who had subtotal hysterectomy in their study, although this surgical method is used in nearly 20% of all benign hysterectomies in Sweden. In the present study, abdominal subtotal hysterectomy and vaginal hysterectomy were independent factors associated with increased patient satisfaction. The explanation of a higher satisfaction rate with subtotal hysterectomy is not obvious. The choice of the surgical approach and technique are always discussed with the patient. Perhaps these patients had a preference for subtotal hysterectomy and therefore experienced a higher satisfaction due to this. On the other hand, subtotal hysterectomy causes a higher risk of persistent vaginal bleeding postoperatively, which could endanger the patient satisfaction 12 . Contrary to Billfeldt et al. who found a significantly higher satisfaction rate after laparoscopic hysterectomy than after abdominal hysterectomy, this study revealed no such advantage of the laparoscopic approach when adjusted for confounders such as complications. Billfeldt et al. presented a multivariable analysis of satisfaction rates, although not adjusted for complications that may have substantial impacts on the outcomes. Even the American study by Pitter et al. did not adjust satisfaction rates for occurrence of complications or other important confounders, although existence of complications seems to be an important significant predictor of dissatisfaction. Thus, none of these studies gave a generalizable answer to the question of satisfaction after various modes of hysterectomy 7 ,  9 . Concerning association of pain as indication of hysterectomy with patient-reported contentment with surgery, Grundström et al. previously reported that women with pelvic pain and endometriosis were at a higher risk of being dissatisfied 5 . Brandsborg et al. stated that 32% of women reported chronic pain after hysterectomy on benign indications. Risk factors for chronic pain were preoperative pelvic pain, pain as main indication for surgery, and pain problems elsewhere preoperatively 13 . This would presumably also influence expected satisfaction with hysterectomy. The negative association between having myoma as indication for the hysterectomy and the perceived medical condition is difficult to explain. The main indication of surgery registered in the GynOp was decided by the responsible gynecologist. Menstrual disorder and pain are symptom diagnoses whereas myoma is a tentative patho-anatomical diagnosis that usually is associated with symptoms such as menstrual disorder or mechanical symptoms. The GynOp does not give norms for how to give priority to the indications. Thus there might be a mixing of symptoms and anatomical abnormalities giving the ultimate main indication. This could possibly influence the results for perceived medical condition. The patientʼs preoperative expectation of surgery is probably also influenced by the preoperative information given, although evidence is lacking. It differs greatly in the beliefs and practices of healthcare professionals in the advice they give to patients 14 ,  15 . Nevertheless, it seems likely that adequate and understandable patient-centered information is important in order to achieve realistic preoperative expectations 16 ,  17 . Patients are more likely to be dissatisfied with result of surgery if they are not satisfied with the preoperative information 18 ,  19 . The association between smoking and the negative outcomes in patient satisfaction after surgery as seen in this study corresponds with findings in other studies 9 ,  20 ,  21 . The reasons are speculative, but psychosocial factors may be of importance. It has previously been suggested that LOS is adversely associated with postoperative complications 22 ,  23 . To the best of our knowledge no previous study has indicated that this association seems to apply even to late appearing complications. Complications that occurred intra- and postoperatively in this study prolonged hospital stay considerably. This is an obvious drawback not only for the exposed women but also considering the deficiency of institutional care and health economics. Interestingly, LOS was even associated with the occurrence and degree of complications reported after discharge. When a major complication occurred after discharge, the LOS was on average a half day longer than for those who had a postoperative period without complications. Although a half day may be considered as clinically insignificant the association seems to indicate that something that happens during the hospital stay may predispose to complications in the postoperative period even up to one year after the surgery. The reasons for this remain speculative and merits further investigation. It therefore seems important to avoid complications not only in the immediate perioperative course but also later to achieve optimal quality of care. This study showed that complications following hysterectomy were strongly associated with less contentment with surgery and an increasing LOS. Prevention of both early and late appearing complications seems to be of importance to optimize patient satisfaction. Other risk factors for adverse outcomes were pain as indication of surgery, having low or no expectations preoperatively, and current smoking. Patient-centered information to ensure realistic preoperative expectations seems essential to gain optimal satisfaction with surgery.

adjusted odds ratio the American Society of Anesthesiologistsʼ classification of physical status body mass index confidence interval the Swedish National Register for Gynecological Surgery length of hospital stay odds ratio standard deviation

This study consists of retrospective data from the Swedish National Register for Gynecological Surgery (GynOp) 8 , on the cohort of women undergoing hysterectomy for benign conditions between January 2004 and June 2016. Exclusion criteria were postmenopausal women or women more than 55 years of age, present adnexal tumor, dysplasia, prolapse or urinary incontinence as main indication for surgery, indication or mode of surgery unclear, preventive or pregnancy-related surgery, and present chronic infectious diseases. More than 75% of women undergoing hysterectomy on benign indications participate in the GynOp 8 . All women received written information about the register before surgery, and could decline to participate if desired. Both women and physicians were involved in the data collection process. The women completed a health declaration form and answered questions about subjective symptoms preoperatively. Preoperative expectation of surgery was asked for and the women had three alternative answers to the question: Get rid of symptoms, Be relieved of symptoms or No expectations of surgery. The gynecologist who performed the preoperative assessment completed a form about preoperative objective findings. The surgeon recorded information about surgery in connection with the operation. The gynecologist filled out a postoperative form concerning the postoperative course at discharge. The day of surgery and discharge from hospital were registered, which enabled calculation of LOS. Eight weeks and one year postoperatively the women received an inquiry questionnaire regarding health status, opinions covering recovery and experiences of surgery, any adverse advents, satisfaction with surgery, and their current medical condition in relation to outcome of surgery. In addition, the gynecologist reviewed and evaluated all inquiry forms to obtain the most correct data 8 . All information from the women and the gynecologistʼs evaluation were recorded in the register. Assembled data included: demographics from health declaration form: age, parity, height and weight, (enables calculation of body mass index [BMI]), smoking habits, and womenʼs preoperative expectations of surgery; clinical data accumulated peri- and postoperatively: the American Society of Anesthesiologistsʼ classification (ASA), main indications of surgery categorized in four groups (menstrual disorder, uterine myoma, pain-related or other indications (premenstrual tension syndrome, hormone therapy related problems other than bleeding disorder, cervical stenosis and hematometra, chronic cervicitis, complications with intrauterine device, and urinary dysfunction related to the uterus), surgical technique (subtotal/total) and approach (at conclusion of surgery, that is abdominal or minimally invasive [laparoscopic or vaginal]), whether the woman had any remaining ovary at conclusion of surgery, complications during hospital stay; from the eight-week inquiry form: occurrence of complications after discharge, and information from the one-year inquiry form covering questions about occurrence of complications from eight weeks postoperatively up to one year, and rating of medical condition at present, and contentment with the result of surgery. The questions (Q) and options for possible answers (A) in the inquiry forms are presented in Fig. 1 . Questions (Q) and options for possible answers (A) in the inquiry forms from the GynOp. The women who underwent surgery less than one year prior to the data retrieval did not have the possibility to answer the one-year questionnaires. Moreover, some of the questions in the inquiries have changed over time. Thus, there will be missing data for some of the items. The question in the preoperative inquiry form concerning the patientʼs expectation of the surgery (get rid of the symptoms; be relieved, or do not have expectations) was excluded in 2013 and the one-year inquiry form concerning medical condition was included 2007 and concerning result of surgery in 2010. The one-year inquiry questionnaire was optional. All except one region in Sweden systematically sent the form to all participating women. The response rate of the one-year inquiry form was consequently determined as the proportion of women answering the inquiry form of those who de facto received the form. The GynOp has classified complications as mild or severe. Severe complications, which are equivalent to major complications, were defined as thromboembolism, all injuries to the bladder, ureter, bowel, or major vascular structures, fistula, bleeding of more than 3000 ml, any reason making a re-operation necessary, hospitalization for more than seven days, persistent physical handicap or death of the patient. Furthermore, septic postoperative infections, and any other major complication (i.e. aspiration, allergic shock, myocardial infarct, or cerebral complication) also constitute severe complications 9 . Mild complications are equivalent to minor complications, and were defined as adverse events that did not have the severity of major complications for example urinary tract infections and wound complications/infections. The software Statistica v 13.2 (Dell Software, 5 Polaris Way, Aliso Viejo, CA 92656, USA) was used for the statistical analyses. Data are presented as mean and one standard deviation (SD) or number and per cent. Continuous data were analyzed by means of one-way ANOVA or Studentʼs t-test, and categorical data by means of χ 2 test for trends. Analyses of LOS in relation to complications were performed by means of ANCOVA models adjusted for BMI, smoking habits, and occurrence of complications developed prior to the occasion of evaluation. Post hoc tests were evaluated by means of Tukeyʼs honest significant difference tests. In order to assess the impact of various factors on outcome measures of PREM and PROM, multivariable logistic regression models were assessed. The dependent variable medical condition was dichotomized in “improved or much improved” versus “unchanged, worse or much worse”. Likewise was the opinion about the result of the surgery one year postoperatively dichotomized in “satisfied or very satisfied” versus “neither satisfied nor dissatisfied”. Multivariable logistic regression models were adjusted for known or potential confounding factors. The confounding factors (age, BMI [as a continuous variable], smoking, indication and mode of hysterectomy and having a remaining ovary postoperatively) were entered simultaneously into the models. Results are presented as crude and adjusted odds ratios (ORs or aORs) and 95% confidence intervals (CI). The significance level was set at p < 0.05. The study was approved by the Regional Ethics Board of Linköping University (Reg. nr. M19–07; amendment 2016/66-32).

Hysterectomy performed on benign indication is the most common major gynecological procedure 1 . Hysterectomy may be accomplished either by removing the entire uterus (total hysterectomy) or retaining the cervical part (subtotal hysterectomy) using an abdominal or minimally invasive (laparoscopic or vaginal) approach 2 . More than 90% of women report a high satisfaction rate several years following hysterectomy, irrespective of surgical approach or technique 1 ,  3 ,  4 . High levels of satisfaction are correlated with improvements in symptoms, such as abnormal uterine bleeding or mechanical symptoms of myoma. Pelvic pain or endometriosis as indication of hysterectomy can be associated with greater dissatisfaction of the result of surgery 5 . However, risk factors for an adverse outcome following hysterectomy in patient-reported measures have rarely been reported in the literature. It seems crucial to focus on achievement of the goals of surgery in a patient perspective, to gain a more patient-centered health care. Moreover, the association between the appearance and degree of intra- and postoperative complications following hysterectomy and the perceived satisfaction rate has not been particularly studied 5 ,  6 ,  7 . Intra- and postoperative complications may occur whatever surgical technique or approach is used, although complications seem less frequent following minimally invasive surgery (around 10% for vaginal hysterectomy, up to 20% reported for abdominal hysterectomy) 3 ,  4 . It seems reasonable to believe that adverse events associated with hysterectomy could affect the womanʼs satisfaction with the result of surgery. The aim of this retrospective register study was to evaluate the impact of intra- and postoperative complications on the perceived satisfaction and self-reported assessment of the medical condition one year after hysterectomy on benign indications, and to determine risk factors for adverse outcomes in these measures. A secondary aim was to analyze whether complications occurring at different times during and after the hospital stay were associated with the length of hospital stay (LOS).