Introduction

Complications are common after surgery.1–5 While major complications after major abdominal surgery are often quoted at around 25%, the incidence of any postoperative adverse event after major abdominal surgery has been identified to be as high as 50%–60%.1 2 A number of factors, in addition to the number of adverse patient events, may impact on documented complication rates. Some of these include definitions used for complications,6 how hard you look for complications,2–5 the period of time in which you look for complications1 7 and the perspective you use when diagnosing complications.1 5 8–10

Checking regularly for complications, and putting hospital systems in place for identifying complications have consistently been shown to increase the number of identified complications,2–5 more than doubling identified complications in some instances.5 The duration of looking for complications is also important, with two studies reporting that one-third of complications are diagnosed after discharge from hospital.1 7

The perspective you use when diagnosing complications is also important. The medical perspective tends to assess outcomes of greatest interest to medical staff, with technical events and defined complications being emphasised. These outcomes are generally emphasised in the process of clinical audit, which traditionally is the main tool used to evaluate surgical outcomes. The patient’s perspective is more focused on the impact that adverse events have on their experience and quality of life (QoL). The emphasis is more on symptoms and provides a more holistic perspective of the postoperative journey. It has been argued that the data obtained from the patient’s perspective on postoperative problems are essential to enable clinicians to comprehensively review the overall success of treatment.8 An example of changing the definition of a complication to include the patient’s perspective is to move from a clearly defined list of complications to a definition, such as: ‘A complication is a condition or an event, unfavourable to the patient’s health, causing irreversible damage, or requiring a change in therapeutic policy, including prolonged hospital stay’.5 An even more inclusive definition is when ‘complications include unexpected events which result in additional patient management problems’.1 As taking a patient-centred approach to identifying complications results in an increase in complications,1 5 9 10 this raises questions about how significant these complications are. Are more ‘minor’, or perhaps inconsequential short-term problems being identified, or are these adverse events significantly impacting on the patient’s quality of life?

A review in 2013 of postoperative complications in general surgery concluded that patient-centred outcomes have not been ‘applied’ when assessing postoperative complications.11 Studies have shown that telephone interviews and questionnaires can be used to gauge patient experience and their QoL.1 12 We had previously developed and validated a set of questionnaires that can complement generic health surveys to prospectively collect information about complications and QoL from the perspective of the patient.13 These questionnaires have been used following cholecystectomy and colorectal surgery.14–17 The first aim of this study was to assess complications after surgery, including after discharge from hospital, using a patient-centred approach. We were interested in identifying what complications or adverse events patients regard as being important and how frequently they experienced these. The second aim was to assess the importance of these events by documenting how they impacted on QoL and the perceived success of their treatment.

Methods

All patients included in the study had undergone surgery in the Department of General Surgery in Dunedin Public Hospital, New Zealand. Using three prospectively collected sources of data: the Otago Clinical Audit, a postdischarge telephone interview and a patient questionnaire, we retrospectively analysed the recovery from surgery of patients operated on between July 2010 and July 2011. One hundred consecutive patients undergoing each of the following types of surgery: hernia repair, biliary surgery, excision of skin lesion, colorectal surgery and anorectal surgery were selected. Exclusion criteria included questionnaires without the telephone interview attached and questionnaires which had not been adequately completed.

The Otago Clinical Audit18 is an established audit programme designed to capture all hospital admissions. All identified surgical complications are entered by the surgical teams shortly after discharge from hospital. The audit includes a list of coded complications, as well as a miscellaneous option which allows the surgical team to include other important events and complications that are not otherwise defined. The audit is then separately checked and signed off by the consultant responsible for the patient’s medical care. The structured telephone interview was performed 2 weeks after discharge by a senior nurse with many years of experience in ward nursing and an active interest in enhanced recovery after surgery. This was designed to identify problems experienced shortly after discharge from hospital. The questions moved from the open-ended question ‘Have you had any problems with your surgery or your recovery?’ to more specific questions about wound problems, infection, excessive bleeding, persistent pain and other. For each issue identified, additional questions were asked to gain more details. These questions were: ‘What exactly was your problem?’, ‘When did it happen and how long did it last for?’, ‘What did you do about it?’, ‘How was it dealt with?’ and ‘Are there any other concerns, comments or suggestions you wish to add?’ The patient questionnaire was sent to patients 4 weeks after discharge and was returned approximately 2 months after the surgery. This was designed to collect information about complications and QoL from the perspective of the patient. The questionnaire included questions used in the 36-item Short Form Health Survey (SF-36 version 2) QoL instrument by Optum, condition-specific questions for different operative procedures, questions about how the patient rated their surgical experience and questions enquiring about problems after surgery. The patient’s perspective on their surgery was assessed by asking them to rate eight questions (figure 1) on a 5-point Likert scale. These enquired about the quality of the information provided, the quality of care, pain after surgery, symptoms and QoL compared with before surgery, how the patient feels about themselves as a result of the surgery, if they would have the surgery again and how happy were they with their surgery. With respect to identifying complications, the questionnaire asked the same questions as those asked in the telephone interview. Questionnaires received back more than 90 days after surgery were excluded to reduce recall error.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

The eight questions asked to assess the patient’s perspective on their surgery and to construct the patient’s satisfaction score.

Results

We studied 100 patients after hernia, biliary, colorectal, skin and anorectal surgery, respectively. The complexity of the surgery covered the range of general surgery including inguinal and abdominal wall hernia repairs, laparoscopic and open cholecystectomy, all types of colonic resections, small bowel resection, local skin excision with some flaps and skin grafts as well as a range of more minor anal procedures including pilonidal surgery, haemorrhoidectomy, fistula and fissure surgery and transanal endoscopic microsurgery. The patients’ mean age was 61 (SD 17.9) years, and 53.8% of respondents were male. Postoperatively, 226 of 500 patients (45.2%) reported at least one complication. There were 344 complications, with 138 of 226 patients (61%) having a complication in one category, 63 (28%) in two categories and 25 (11%) in three or more categories. As complications could be reported at audit, telephone interview or questionnaire, a total of 411 events were reported, with 16% of complications being reported on more than one occasion.

In terms of the three time periods used to capture complications, the breakdown for all reported complications is summarised in table 1.

In table 2, 411 reported events are summarised according to the operative procedure performed, and also the timing at which the event was reported.

Of the 411 events, 12.6% were captured by the audit, 36.3% by telephone and 51% by questionnaire. The most frequent categories of complications identified (table 2) at the audit were wound problems and ‘other’, at the 2-week postdischarge telephone call was pain and at 2 months by the questionnaire were wound, then pain and then other.

The identification of patients who developed a complication, and when this was first diagnosed, is summarised in table 3.

Of the 226 patients who developed complications, 45 (20%), 91 (40%) and 90 (40%) were identified for the first time by the audit, telephone interview and postal questionnaire, respectively. The risk of a patient developing a complication was 57%, 56%, 39%, 38% and 36% after colorectal, anorectal surgery, hernia surgery, biliary surgery and skin surgery, respectively. 26% of anorectal, 25% of colorectal, 18% of biliary, 13% of hernia and 6% of skin operations developed more than one complication. Patients having colorectal surgery were most likely to be diagnosed with a complication before discharge from hospital (audit identified 42% of colorectal complications) and those with anorectal surgery were least likely to be diagnosed before discharge from hospital (audit identified 5% of complications). Patients having anorectal surgery were most likely to be identified as having a complication for the first time at 2 weeks after discharge (61% of anorectal complications). Patients having skin surgery were most likely to be identified as having a complication for the first time at 2 months after surgery (58% of skin complications), mainly because of late presentation of wound infections.

The results for the 344 complications are summarised in table 4.

The order of complications by category, from most to least frequent, was wound, pain, other, infection and bleeding. Wound complications were present in 24% of patients and accounted for 35% of complications. Wound infection was the main contributor across all three stages (table 1) and was most likely to be identified after discharge from hospital. Pain issues were present in 20% of patients and accounted for 29% of complications. Pain was rarely identified as a problem by doctors in the audit but was often identified as a major problem by patients after discharge from hospital, especially after anorectal procedures. ‘Other’ included patients with medically serious complications such as acute renal failure, pulmonary embolism, cardiac arrhythmia and congestive heart failure, which were usually captured by the audit. However, the majority of problems in the ‘other’ category were functional gastrointestinal problems such as constipation, diarrhoea and stoma problems. These functional gastrointestinal problems were usually identified after discharge from hospital.

The impact of complications on how the patient perceived their surgical experience is summarised in figure 2. There are two main findings. As patients without and with complications all received good-quality medical care, overall there was a good level of satisfaction with the care received. The mean difference in patients with complications was usually one position lower along the Likert scale, resulting in a change of overall rating from ‘excellent’ to ‘very good’—or from ‘a lot better’ to ‘somewhat better’. When all patients with complications were combined into one group, they continued to be very happy with the care they received and based on their experience would agree to have the surgery again. However, the patient feedback also highlighted a number of important concerns. These concerns are summarised by the specific answers given, by differences in the patient satisfaction score and by the frequency of responses given in the worst two options on the Likert scale. Whenever there was a complication (compared with patients who did not have a complication), patients did not believe that the information they had received about the procedure was as good. Patients with complications had more postoperative pain, the improvement in how they felt about themselves was less, their improvement in QoL on direct questioning was less, and their overall satisfaction with the surgery was lower. In addition to this, on further breakdown by complication type, patients with a wound complication or an infection felt that the care they received was not as good, and those with infection or bleeding were less likely, based on their experience, to undergo the same operation again. All the comparisons, comparing the ratings between patients with and without complications, were significantly different, p<0.001. The patient satisfaction score, which gives a more quantitative overview of the patients’ rating, was 85.6 (11.2) (mean (SD)) for those with no complications and 74.6 (18.9) for those with complications, p<0.001. Another way to examine patient ratings is to look the frequency of scores made in the worse two positions on the Likert scale. This was always statistically significantly different between patients with and without complications. For patients with complications, the frequency ratings in the worse two positions were approximately 5% for quality of information given, care received and how happy the patient was with their surgery; were approximately 10% for rating of symptoms compared with before surgery, would you have the surgery again and how the patient felt about themselves; were approximately 15% for the individuals rating of their QoL and were 37% for pain.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

Median survey ratings on the Likert scale for patients with and without complications according to complication type. Error bars: 95% CI. A: No complication, B: All patients with complications, C: Wound, D: Infection, E: Bleeding, F: Pain, G: Other. The distribution for all results comparing patients with any complication against patients without complications using the Cochran-Armitage test was consistently significantly different with a p value of <0.001. Additional details on the ratings on the individual Y scales are presented in figure 1.

The overall QoL as reported by SF-36 was significantly different for those with and without complications. For the PCS, this was 48.5 (9.2) (mean (SD)) for those without complications and 43.9 (10.2) for those with complications, p=0.021. For the MCS, this was 51.2 (9.2) and 47.6 (10.1), respectively, p=0.055. A multiple linear regression analysis with the SF-36 PCS score as the dependent variable showed that complications, age and ASA were all similarly predictive, with a significance of p<0.001 and a standardised beta coefficient of 0.19 for complications, 0.20 for age and 0.21 for ASA. Multiple linear regression analysis with the SF-36 MCS as the dependent variable showed that complications made the greatest contribution to the MCS, with a standardised beta coefficient of 0.17, p<0.001. In comparison, the beta coefficient and p value for age was 0.10, p=0.07, and for ASA was 0.11, p=0.05. Timing of surgery and sex did not contribute to either the PCS or MCS. The Pearson correlation coefficient between the patient satisfaction score and PCS was 0.348 and between the patient satisfaction score and the MCS was 0.406.

Discussion

The main research findings of our study are that patients commonly experience problems such as wound infection, pain and functional gastrointestinal symptoms which are often not identified by conventional surgical audit. This may be because these events develop after discharge from hospital, or because audit would not classify these events as complications if they were observed. Although this study is not designed to distinguish between complications which may or may not have been identified using standard audit definitions, these patient-reported complications are important because they are associated with a reduced QoL, a reduced satisfaction with surgery and a worse rating of the patients’ postoperative course.

For a study involving a spectrum of minor and major general surgical operations, the frequency of complications was high, with 42.5% of patients developing a complication. Using clinical audit alone we would have identified 20% of the patients who developed a complication. The fact that the telephone follow-up at 2 weeks and the questionnaire at 2 months identified approximately 80% of patients with complications illustrate the importance of directly contacting the patient. This also highlights two issues with respect to identifying complications. The first is that audit systems, where inpatient complications are identified by medical staff, only capture a small proportion of the number of events that trouble patients.4 5 21 This does not minimise the importance of a medically led audit, but it does remind us that this usually represents only part of the patient’s journey. The second issue is related to the timing of when complications develop. As a number of patients had operations with a short hospital stay, we would have expected a significant proportion of postoperative problems to develop after discharge from hospital. The observation that 80% of patients who developed complications were initially identified after discharge from hospital is much higher than the 33% previously identified in other studies.1 7 Although this difference is partly explained by differences in studies, such as more minor procedures and an earlier discharge in our study, this result also emphasises the importance of ongoing patient surveillance after discharge.

The most common problems identified after discharge from hospital included wound problems (especially wound infection), pain and ‘other’ functional problems. The majority of wound infections being diagnosed after discharge from hospital is consistent with what has previously been documented in the literature,22 23 with different studies demonstrating that only 50%–80% of infections are identified by the postoperative day 16.24–26 With respect to pain, 12% of patients identified this as a problem at 2 weeks after discharge, and 10% were still experiencing problems 2 months after surgery. This is in marked contrast to the medical audit. This difference in identification of pain by audit and by patient reporting highlights the importance of the perspective of the person reporting the complication. This result also suggests that the management of pain is often inadequate when moving from the hospital environment to the home environment, especially after anorectal procedures. The frequency of functional problems (nausea, vomiting, constipation, stoma leakage) in the ‘other diagnosis’ category, which was present in over 5% of patients after discharge, has also been previously noted1 and is another area where a more active management strategy is required after discharge from hospital. These results raise a number of key issues. The first is that in our study, when attempting to identify complications, the use of clinical audit at discharge under-reports complication rates.4 21 We need to be able to recognise a broader set of outcomes than those identified by clinical audit and we need to have a more robust strategy for documenting complications that can develop over the weeks following discharge from hospital. The second is that hospitals need to have better systems of support in place for patients after they are discharged. We believe that the greater emphasis on early discharge from hospital needs to be matched with an equal emphasis on improving the quality of care immediately after discharge. Third, in terms of transparency, we need to have a better knowledge of the postoperative problems experienced by patients after discharge from hospital, so that we can give our patients correct information when they are consented and provide appropriate expectations about the difficulties they may face.

Perhaps, the main advantage of the postoperative questionnaire was that this enabled us to assess the impact complications had on the patient’s postoperative journey. When comparing patients who developed complications against those who did not develop complications, a number of differences were noted. First, patients with complications did not feel that they had been as well informed about their surgery, with the rating of information they received decreasing from ‘excellent’ to ‘more than adequate’. This highlights issues around discussing potential problems before surgery, including making sure patients have an appropriate understanding about what they can expect to happen after their surgery. Second, patients with complications consistently experienced more pain, which was often still a problem 2 months after surgery. Third, on direct questioning, the improvement in QoL was reduced from ‘a lot better’ in those without complications to ‘a bit better’ when compared with before their surgery. There was also a significant lower SF-36 PCS (p=0.021) and an almost signficantly lower MCS (p=0.055) in patients with complications. These results highlight that postoperative complications result in a medium-term impact on surgical recovery. Lingering symptoms and a slower recovery mean that 2 months after surgery the patient continues to experience an impaired QoL. In the literature, a lower QoL after complications and after wound infection has previously been noted.27 28 Our study takes this observation further by demonstrating that when postoperative complications include problems identified by the patient that there continues to be a significant reduction in their QoL. Our linear regression analysis also confirmed that the magnitude of the impact complications has on QoL is similar to the impact that age and comorbidities (as measured by the ASA score) have on the PCS component of QoL. Fourth, for patients with complications, there was also a reduction in satisfaction about surgery from ‘extremely happy’ to ‘quite a bit’ happy. The overall patient satisfaction score was also significantly reduced. It is recognised that an inherent limitation of asking about patient satisfaction is that people tend to be quite satisfied with the care they received. Our patient satisfaction score is drawn from the eight questions (figure 1), which reflect a mix of patient experience as well as satisfaction. This was significantly different in patients with complications and also had an excellent correlation to the SF-36 PCS and MCS scores. These results support further work examining the assumptions underlying the patient satisfaction score, as well as the validity and potential uses of this score.

The questionnaire also identified some specific issues after wound problems, infective problems and bleeding. Patients with wound or infection problems felt that the care they had received was not as good. This was an interesting finding as it implies that an infection is perceived to be ‘at least partly’ preventable. Although individual cases of wound infection may not be preventable, our patients’ perception is supported by evidence in the medical literature that introducing ‘package of care’ programmes, which include improving compliance with best practices, can reduce infective complications.29 Patients with infection problems or bleeding were also less likely to be willing to repeat their surgery, although the reasons for this finding were not clear.

While some of these findings would be predictable, the confirmation that the improvement in QoL is less in patients with a range of complications 2 months after surgery, as well as a decreased rating about the quality of information received and the quality of their postoperative recovery is important. Both the frequency of postoperative problems and the patient’s feedback reveal gaps and frustrations with postdischarge care, which may negatively impact on clinical outcomes and impact on their QoL.28–30

While the finding that scoring complications from the patient’s perspective increases the number of identified complications identified1 5 9 10 is again confirmed, our study provides additional qualitative and quantitative data about these complications. These complications were not minor or ‘inconsequential’. They were clearly of significance to the patient, resulting in differences in QoL and ongoing morbidity for at least 2 months following their surgery. The observation that most patients report problems that are personally and clinically significant is similar to a study of complications after back surgery which demonstrated that 50% of patient-reported problems were still producing significant symptoms and difficulties 1 year following surgery.31 In this context, it should be argued that patient reporting would improve our appreciation of ‘real’ postoperative complication rates.21

The retrospective nature of our analysis resulted in our study having a number of limitations. Although the questionnaire we used had been previously validated, we cannot guarantee the accuracy of all the information received as we were unable (in real time) to clarify responses to some of the questions or to independently confirm complications. The reliability of information could also be influenced by recall bias (eg, forgetting something that was a problem a month ago) or from the under-reporting or over-reporting of symptoms and patients not correctly understanding all the questions they were answering. However, this study does demonstrate that useful information can be obtained by the use of questionnaires and would support the routine use of questionnaires for capturing complications following discharge from hospital. In terms of analysis, while both time after surgery and the perspective of the patient were important in identifying complications after discharge, we were unable to quantify their individual impact on the overall diagnosis of complications. This would require a prospective study.

In conclusion, our study demonstrates the high rate of problems patients experience after a range of surgical procedures. These problems often develop after discharge from hospital and have an ongoing impact on the patient’s QoL and satisfaction over a period of at least 2 months. One advantage of taking a patient-centred approach to documenting postoperative problems is that it does help to highlight system problems where improvements in care can be delivered.