Abstract
Aim:
It has been reported in the literature that patients with poor preoperative mental health are more likely to have worse functional outcomes following primary total hip and knee arthroplasty. We could find no studies investigating whether preoperative mental health also affects length of hospital stay following surgery. The aim of this study was to determine whether preoperative mental health affects length of hospital stay and long-term functional outcomes following primary total hip and knee arthroplasty. We also aimed to determine whether mental health scores improve after arthroplasty surgery and, finally, we looked specifically at a subgroup of patients with diagnosed mental illness to determine whether this affects length of hospital stay and functional outcomes after surgery.
Method:
Through a review of prospectively collected regional joint registry data, we compared preoperative mental health scores (SF-12 MH) with length of hospital stay and post-operative (1 and 5 years) functional outcome scores (Oxford and Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC)) in 2279 primary total hip and knee arthroplasty surgeries performed in the Bay of Plenty District Health Board between 2006 and 2010.
Results:
Based on Pearson product–moment correlation coefficients, there was a significant correlation between preoperative mental health scores and post-operative Oxford scores at 1 year as well as post-operative WOMAC scores at both 1 and 5 years. There was no significant correlation between preoperative mental health and length of hospital stay. Mental health scores improved significantly after arthroplasty surgery. Those patients with a formally diagnosed mental illness had significantly worse preoperative mental health and function scores. Following surgery, they had longer hospital stays although their improvement in function was not significantly different to those without mental illness.
Conclusion:
The results of this study support reports in the literature that there is a correlation between preoperative mental health and long-term functional outcomes following primary total hip and knee arthroplasty. Patients with poor preoperative mental health are more likely to have worse functional outcomes at 1 and 5 years following surgery. No correlation between preoperative mental health and length of hospital stay was identified. Mental health scores improved significantly after surgery. Patients with mental illness had longer hospital stays and despite worse preoperative mental health and function had equal improvements in functional outcomes.
Keywords
Introduction
Arthritis of the hip and knee are common causes of pain and functional impairment. The relief of pain and restoration of function are the main goals of primary total hip and knee arthroplasty. Arthroplasty surgery is the definitive end-stage treatment for hip and knee arthritis for patients who have exhausted all non-operative measures. Both hip and knee arthroplasty surgeries have been shown to be highly successful in improving functional outcomes. 1,2
It has been reported in the literature that patients with poor preoperative mental health are more likely to have worse functional outcomes following primary total hip and knee arthroplasty at 1 and 2 years post-operatively. 3,4 However, despite these findings, no published studies have reported that poor mental health results in prolonged length of hospital stay following surgery.
The aim of this study was to determine whether preoperative mental health affects length of hospital stay and long-term functional outcomes following primary total hip and knee arthroplasty. We also aimed to determine whether mental health scores improve after surgery. Finally, we looked specifically at a subgroup of patients with a diagnosed mental illness to determine whether this affects length of hospital stay and post-operative functional outcomes.
Method
After obtaining Ethical Review Board approval, data covering the period from January 1, 2006 to December 31, 2010 were collected from our regional joint registry records. Our regional joint registry was established in 2004 and collects prospective data on a wide range of orthopaedic conditions. This registry was established by a group of orthopaedic surgeons who serve a population of approximately 250,000 people. Our regional joint registry prospectively records preoperative functional scores, baseline demographics, operative characteristics and post-operative outcome measures for all patients undergoing arthroplasty surgery throughout the region. The registry assesses preoperative and post-operative patient function using self-administered disease specific (Oxford and Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC)) and general health (SF-12 PH and SF-12 MH) questionnaires completed by patients at their preoperative clinic appointment as well as at 1 and 5 years post-operatively.
Registry records of all patients who underwent primary total hip and knee arthroplasty during this 5-year study period were reviewed. Those who had undergone simultaneous bilateral procedures were excluded from the study.
Baseline demographics (age at time of surgery, gender, American Society of Anesthesiologists (ASA) score, operative type, operative side, preoperative diagnosis and any formally diagnosed mental illness) were recorded as well as length of hospital stay following surgery from patient medical records.
Preoperative mental health scores (SF-12 MH) were compared with length of hospital stay and 1- and 5-year post-operative functional outcome scores (Oxford and WOMAC). Preoperative mental health scores were also compared with post-operative mental health scores. Finally, we compared those with a formally diagnosed mental illness to those without to determine whether there was any difference in length of hospital stay and functional outcomes.
Statistical analyses
Standard descriptive statistics including means, standard deviations, frequencies and percentages were used to summarize the continuous preoperative and post-operative measures as appropriate. Pearson’s product–moment correlation coefficients were used to measure the association between preoperative mental health and length of hospital stay, post-operative functional outcomes and functional outcome changes. Paired t-tests were used to test the changes in mental health scores and independent t-tests to compare the outcomes between preoperative mental health diagnosis groups and to compare preoperative mental health between the outcome groups. A two-tailed p-value of <0.05 was taken to indicate statistical significance.
Results
From our regional joint registry records, we found that there had been 2389 primary total hip and knee arthroplasty surgeries performed during this 5-year study period with a minimum of 1-year follow-up. Those patients who had undergone simultaneous bilateral arthroplasties were excluded leaving a total of 2279 patients to be included in the study. Not all measures were available for all patients at each time point and mean levels and sample sizes for each measure at each time point are shown in Table 1.
Preoperative mental health scores (SF-12 MH), length of hospital stay (days) and 1-and 5-year post-operative Oxford and WOMAC scores.
WOMAC: Western Ontario and McMaster Universities Osteoarthritis Index.
At the time of surgery, the average age of all patients was 69.18 year and the average ASA score was 1.99. There were 1053 male (564 primary total hip arthroplasties and 489 primary total knee arthroplasties) and 1226 female patients (656 primary total hip arthroplasties and 570 primary total knee arthroplasties). There were 1220 primary total hip arthroplasties and 1059 primary total knee arthroplasties in the study.
The Pearson correlation coefficients indicated that there was a significant correlation between preoperative mental health and three of the four post-operative functional outcome scores but not between preoperative mental health and length of hospital stay (Table 2).
Comparison of preoperative mental health scores (SF-12 MH) with post-operative functional outcome scores (Oxford and WOMAC scores at 1 and 5 years) and length of hospital stay (days).
The Pearson correlation coefficient (r) for preoperative mental health scores and post-operative Oxford scores at 1 year was significant at 0.146 (p < 0.01). This means that as preoperative SF-12 MH scores increase (better mental health), so too do 1-year post-operative Oxford scores (better functional outcome). Although the trend remained, this correlation was not significant for preoperative mental health scores and post-operative Oxford scores at 5 years (r = 0.101, p = 0.059). The Pearson correlation coefficient for preoperative mental health scores and post-operative WOMAC scores was significant at both one (r = −0.148, p < 0.01) and 5 years post-operatively (r = −0.118, p < 0.05). Further analysis of the categories within the WOMAC score showed a significant association between preoperative mental health and 1-year post-operative pain (r = −0.140, p < 0.01), stiffness (r = −0.123, p < 0.01) and physical function (r = −0.145, p < 0.01) as well as 5-year stiffness (r = −0.130, p < 0.05) and physical function (r = −0.124, p < 0.05) although not for pain (r = −0.080, p = 0.139). This indicates that as preoperative SF-12 MH scores increase (better mental health), WOMAC scores decrease (better functional outcome). The Pearson correlation coefficient for preoperative mental health scores and length of hospital stay was 0.013 (p = 0.548) indicating no significant correlation. Table 2 summarizes these findings.
When we compared preoperative SF-12 scores (mean 38.002, SD 12.52) with post-operative scores (mean 52.005 at 1 year, SD 10.36, and mean 51.592 at 5 years, SD 9.82), we found that they improved significantly after arthroplasty (p < 0.01). Table 3 summarizes these findings.
Comparison of preoperative mental health scores with 1- and 5-year post-operative mental health scores.
From our study population of 2279 patients, we identified 66 patients who had a diagnosed mental health disorder coded as part of their medical records (depression, major depressive disorder, anxiety, schizophrenia, schizoaffective disorder, bipolar disorder, or psychosis). When we compared these patients to those without such a diagnosis, we found that, as would be expected, those with mental illness had significantly worse preoperative mental health scores (27.979 vs. 37.761, p < 0.01). They had significantly longer hospital stays (7.682 days vs. 6.653 days, p < 0.01) but their improvement in function from preoperative to post-operative was equivalent (preoperative to 1-year post-operative improvement in Oxford score 27.64 vs. 26.53, p = 0.447, preoperative to 5-year post-operative improvement in Oxford score 26.455 vs. 26.326, p = 0.968, preoperative to 1-year post-operative improvement in WOMAC score 56.554 vs. 54.175, p = 0.465 and preoperative to 5-year post-operative improvement in WOMAC score 55.418 vs. 53.541, p = 0.787). Table 4 summarizes these findings.
Comparison of preoperative and post-operative scores and length of hospital stay in patients with a diagnosed mental illness versus those without.
WOMAC: Western Ontario and McMaster Universities Osteoarthritis Index; ANOVA: analysis of variance.
Discussion
There are a number of studies in the literature which have reported that patients with poor preoperative mental health are more likely to have worse functional outcomes following primary total hip and knee arthroplasty. 3,4 However, there is a paucity of data investigating whether preoperative mental health also affects length of hospital stay following surgery. Similar to other studies, our study clearly demonstrates a significant correlation between preoperative mental health and post-operative function. We did not identify any correlation between preoperative mental health and length of hospital stay following surgery. Patients with poor preoperative mental health are more likely to have poorer functional outcomes at 1 and 5 years following surgery. Conversely, patients with better preoperative mental health are more likely to have better functional outcomes at 1 and 5 years following surgery. Patients in our study population had significant improvements in their mental health following surgery. When we looked specifically at those patients with a diagnosed mental illness, they had significantly worse preoperative mental health and function scores, following surgery their duration of hospital stay was significantly longer but their overall improvement in function was no different from those without mental illness.
Methodological considerations
The SF-12 MH score is a validated scoring tool which was developed using 12 items from the Medical Outcomes Study 36-Item Short-Form Health Surgery (SF-36), a self-administered questionnaire measuring general health status and quality of life with high reliability and consistency. 5 The SF-12 MH score has been shown to accurately and efficiently reproduce the Mental Component Summary scale of the SF-36 with good sensitivity and specificity where a higher score indicates improved mental health status. 6 It can be used as a screening tool for depressive disorders, for monitoring the prevalence of mental disorders in a population and for targeting new treatments as accurately as other tools developed specifically for the detection of depressive disorders. 7 Although used for the screening of any mental disorder, it has only been validated for the use in anxiety and depressive disorders (with a greater ability to detect depressive disorders than anxiety disorders). 7 The Oxford hip and knee scores are instruments for assessing hip and knee function and have both been validated in several studies. 8 The greater the Oxford score, the greater the function. The WOMAC score is a validated scoring system specific to locomotor capability that is widely used to evaluate patients undergoing total hip and knee arthroplasty surgery. 9 The WOMAC is composed of three sections—pain, stiffness and physical function, with a higher score corresponding to poorer function. Although we employed several different functional measures, there are some intrinsic limitations to the use of the Oxford, WOMAC and SF-12 MH scores, which may not be able to measure subtle changes in function.
Our study has certain limitations. Patient numbers are relatively small and follow-up relatively short with post-operative outcome scores recorded at only 1 and 5 years following surgery.
Outcomes
This study has shown a clear correlation between preoperative mental health status and post-operative functional outcomes with a significant correlation between preoperative mental health and 1-year post-operative Oxford scores as well as 1- and 5-year post-operative WOMAC scores. Although this correlation did not exist between preoperative mental health and 5-year post-operative Oxford scores, likely due to the fact that the number of returned 5-year post-operative Oxford questionnaires was small with only those patients who had undergone surgery in 2006 or 2007 having completed 5 years of follow-up, the trend remained for a worse functional outcome in patients with poorer preoperative mental health.
This study has shown no correlation between preoperative mental health status and length of hospital stay following surgery. This was surprising as we postulated that those patients with poor preoperative mental health would have a prolonged length of hospital stay following surgery. This study took place in a District Health Board consisting of two distinct hospitals. The smaller of the two, which serves a population of approximately 50,000 people, runs a multidisciplinary preoperative clinic involving input from Physiotherapists, Occupational Therapists and Social Workers. The aims of this clinic are to provide patients with preoperative counseling, to prepare them for surgery and to educate them about what to expect post-operatively. The average length of hospital stay in this hospital (based on data from the previous 12 months) is 3.53 days following arthroplasty surgery. In the larger of the two hospitals, serving a population of approximately 200,000, no such multidisciplinary preoperative clinic exists and the average length of hospital stay for arthroplasty patients is 5.78 days. Therefore, although our study has shown no clear correlation between preoperative mental health and duration of hospital stay, this is something that deserves further research as there is a definite difference in average length of hospital stay following arthroplasty surgery between our two hospitals which may well be a result of this preoperative multidisciplinary clinic and its influence on patient expectations around surgery.
Among our study population mental health scores improved significantly after arthroplasty surgery. This could potentially indicate that a component of a patient’s mental distress may be attributable to their orthopaedic condition and that by addressing this, through arthroplasty surgery, can lead to improvements in mental health. For example, it could be postulated that preoperative pain leads to poor mental health and improving pain, which arthroplasty reliably does, can result in improved mental health.
The minimum clinically important difference (MCID) for the SF-36 questionnaire is five points and it has been shown to be able to detect changes in health status as quickly as 3 months after arthroplasty although it is to be seen whether this can be transferred to the SF-12 questionnaire. 10 In a study examining outcomes after anterior cervical discectomy and fusion, the MCID for the SF-12 MH questionnaire was 4.7 points. 11 If this result is applicable in the setting of arthroplasty, then the improvements in patient mental health scores following surgery are not only statistically significant but represent perceivable real life differences also.
Although the 66 patients with a diagnosed mental illness had worse preoperative mental health, worse preoperative function scores and their duration of hospital stay following surgery was significantly longer than those without a diagnosed mental illness their actual improvement in function was not significantly different. This finding would support the fact that patients with mental illness should not be discriminated against when being assessed for consideration of arthroplasty surgery and should have equal access to surgery as those without mental illness (although with additional input from mental health services to optimize preoperative mental health status before embarking on surgery). These results do suggest a significant under-reporting of mental illness on our hospital’s coding database though as the incidence of mental illness amongst adult New Zealanders has been reported as 16% compared with just 2.9% (66/2279) in our study population which could therefore impact on these results. 12
Williams et al. reported that following total hip arthroplasty the mean Oxford score can be expected to increase by an average of 20 points and 14 points following total knee arthroplasty. 13 When we separated our study patients into those whose Oxford scores improved as would be expected preoperatively to 1-year post-operatively versus those whose scores did not, we found no significant difference in preoperative mental health scores indicating that even those patients with poor preoperative mental health can be expected to have improvements in function equal to those with better preoperative mental health (see Table 5).
Comparison of preoperative mental health scores in those whose Oxford scores improved as would be expected versus those whose scores did not.
ANOVA: analysis of variance.
Kalairajah et al. have recommended a category of excellence for an Oxford score of >41, good for a score of 34–41, fair for 27–33 and poor for those <27. 14 Our goal is for all patients to achieve an excellent or good outcome. When we separated our study patients into those who achieved an excellent or good outcome based on their 1-year post-operative Oxford scores versus those who achieved only a fair or poor outcome, we found a significant difference in their preoperative mental health scores (39.020 vs. 35.555, p < 0.01; see Table 6). This result provides support for optimizing a patient’s preoperative mental health status before proceeding with arthroplasty surgery. It is important to identify those patients who are likely to have poor outcomes following surgery as these patients pose a significant cost to the health system. Being able to identify patients with worse preoperative mental health may help surgeons to provide better preoperative counselling or improved access to information to help these patients understand the proposed surgery and the post-operative rehabilitation in an effort to help them set realistic expectations and thus improve their post-operative outcomes.
Comparison of preoperative mental health scores in those who achieved an excellent or good outcome based on their 1-year post-operative Oxford scores versus those who achieved only a fair or poor outcome.
ANOVA: analysis of variance.
Our goal is for all of our patients to achieve an excellent or good outcome following arthroplasty surgery (i.e. a 1-year post-operative Oxford score of >34 according to Kalairajah 14 ). However, a more realistic expectation is for 75% of patients to achieve this. From the results of our study, if a patient has a preoperative SF-12 MH score of at least 30, then their likelihood of achieving an excellent or good outcome is 75% (see Table 7). We therefore recommend input from mental health services for any patient with a preoperative SF-12 MH score of less than 30 in order to optimize their mental health status before proceeding with arthroplasty so as to achieve the best possible post-operative outcome. Given that all patients are required to complete the SF-12 MH questionnaire at their preoperative clinic appointment, this is achievable simply by screening the results and referring patients on for further management by mental health services if necessary.
Receiver Operating Characteristic (ROC) curve analysis using preoperative SF-12 MH scores to predict post-operative outcome (Oxford scores at 1 year).
Although most patients have improved functional outcomes following primary total hip and knee arthroplasty, it has been reported that between 15% and 30% report little or no improvement. 15,16 In some cases, this can be attributed to ongoing pain, restricted range of motion, poor function, or a surgical complication; however, in most cases, no specific reasons can be identified. And it is here that it has been postulated that poor preoperative mental health, psychological distress or depression may be an important predictor of poor post-operative outcome. It has even been reported that preoperative mental health status is the strongest determinant of functional outcome following hip and knee surgery. 17,18 Lingard et al. reported that patients with low mental health scores and other co-morbid conditions before total knee arthroplasty are more likely to have worse outcomes at 1 and 2 years post-operatively. 3 Another study by Lavernia et al. reported that low baseline mental health score was one of the strongest predictors of a worse post-operative WOMAC score. 4 In Lavernia’s study, 26% of the patient population were mentally distressed at the time of surgery (as defined by a SF-36 MH score of less than 52). 4 These patients had significantly worse scores on measures of well-being, pain, stiffness and function post-operatively as well as worse WOMAC scores. 4 They noted, however, that although these patients had worse preoperative baseline measures, their overall improvement following surgery was similar to the non-distressed patients which we also found in our study. 4 An interesting finding in their study, which we also found in ours, was that arthroplasty surgery actually resulted in a reduction in the prevalence of psychological distress (from 26% to 9% in their patient population) likely as a result of a reduction in physical disability associated with arthritis. 4
This relationship between pain and disability, caused by arthritis, and mental health status is a complex one and although we do not suggest that a patient’s preoperative mental health status should in anyway influence their access to surgery, it is important that surgeons are aware of this association. Only by being aware of this link can surgeons provide patients with counselling and modify the way in which these patients are managed in order to ensure that optimal post-operative outcomes are achieved. Providing patients who have poor preoperative mental health with counselling and input from mental health services both before and after surgery may potentially improve their mental health status, thereby also improving their functional outcomes following arthroplasty. This may involve at a minimum a more in-depth discussion between the surgeon and patient or potentially even input from psychologists or other mental health service providers. It is important to note, however, that a low preoperative mental health score does not in itself indicate clinically significant psychiatric pathology and is simply a general measure for a patient’s overall mental health status.
Conclusion
There is no doubt that preoperative mental health and preoperative physical function are intimately linked. However, it is difficult to say whether poor mental health causes poor function or whether poor functional status leads to poor mental health. Either way, our study has confirmed the findings of other reports that a poor mental health score preoperatively is significantly associated with a poor post-operative functional outcome score. This relationship between mental health and functional outcome is a complex one where surgery can positively influence mental health status, while at the same time, poor preoperative mental health status can negatively affect post-operative functional outcomes. A brief screening of mental health status at the time of orthopaedic consultation is worthwhile to identify patients with poor preoperative mental health and instigate appropriate treatment in order to achieve better post-operative functional outcomes.
Footnotes
Declaration of conflicting interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
