Psychological distress and contentment after primary total knee replacement
To my beloved wife Sija, my angels Emir, Dinah, and Talya, my parents, and my brothers, sisters, relatives, and friends
"If you desire the present life, seek knowledge; if you desire the hereafter, seek knowledge; if you desire both combined, seek knowledge too."
Örebro Studies in Medicine 217
A
AMIRS
ALAHM
AHDIPsychological distress and contentment after primary
total knee replacement
©
Aamir Salah Mahdi, 2020
Title: Psychological distress and contentment after primary total knee replacement. Publisher: Örebro University 2020
www.oru.se/publikationer Print: Örebro University, Repro 08/2020
ISSN1652-4063 ISBN978-91-7529-346-2
Abstract
Aamir S. Mahdi (2020): Psychological distress and contentment after primary total knee replacement. Örebro Studies in Medicine 217.
The overall aim of this thesis was to study patients with anxiety/depression symptoms before and after primary total knee replacement, particularly in terms of improvements in patients-reported outcomes, and particularly among those who were discontent with the surgery. This could hopefully improve patients' contentment after TKR.
Study I investigated the prevalence of knee symptom improvement among patients with preoperative anxiety and/or depression in comparison to patients who did not have anxiety/depression. Study II investigated changes in the prevalence of anxiety and depression one year after primary TKR. Study III investigated how TKR surgeons in Sweden informed their patients preoperatively, and what kind of information they gave. Study IV used face-to-face interviews to capture experiences of discontentment one year after TKR among patients without documented complications.
The main findings were that: 1. All patients improved in outcome measures one year after TKR, regardless of presurgical psychological state. 2. Among the 15% of patients with anxiety symptoms before surgery, 59% had improved in these symptoms one year after surgery; while among the 10% with depression symptoms before surgery, 60% had improved one year after surgery. 3. Knee surgeons in Sweden have considerable awareness of the importance of preoperative patient information, the impact of pa-tient expectations, and psychiatric illness. However, they need to improve their preoperative routines when it comes to providing written information, asking about the patient’s expectations, and psychiatric assessment. 4. The patients experienced unfulfilled expectations and needs regarding unresolved and new problems, limited independence, and lacking relational support.
Keywords: Anxiety, contentment, depression, discontentment, psychologi-cal distress, patients’ expectations, total knee replacement.
Aamir S. Mahdi, School of Medical Science. Faculty of Medicine and Health, Örebro University, SE-701 82 Örebro, Sweden,
Table of Contents
LIST OF PAPERS ... 9
ABBREVIATIONS AND DEFINITIONS ... 10
1 INTRODUCTION ... 11
2 BACKGROUND ... 12
2.1 Osteoarthritis ... 12
2.2 Total knee replacement ... 13
2.3 The Swedish Knee Arthroplasty Register ... 17
2.4 Patient-reported outcome measures ... 18
2.4.1 Knee Injury and Osteoarthritis Outcome Score ... 19
2.4.2 EQ-5D-3L ... 20
2.4.3 Visual analogue scale ... 21
2.4.4 Hospital Anxiety and Depression Scale ... 22
2.5 Contentment after TKR ... 23
2.6 Rationale ... 24
3 AIMS OF THE THESIS ... 25
4 PATIENTS AND METHODS... 26
4.1 Settings ... 26 4.2 Samples ... 27 4.3 Procedures ... 32 4.4 Outcome measures ... 33 4.5 Demographic variables ... 37 5 ANALYSIS ... 39 5.1 Study I ... 39 5.2 Study II ... 40 5.3 Study III ... 41 5.4 Study IV ... 41 6 ETHICAL CONSIDERATIONS ... 43 7 SUMMARY OF RESULTS ... 44 7.1 Study I ... 44 7.2 Study II ... 51 7.3 Study III ... 58
8 GENERAL DISCUSSION ... 66
8.1 Symptom improvement after TKR ... 67
8.2 Anxiety/depression prevalence after TKR ... 69
8.3 TKR in Sweden ... 70
8.4 Patients’ experience of discontentment after TKR ... 72
8.5 Methodological considerations ... 74
8.6 Implications for future research ... 78
9 CONCLUSIONS ... 80
10 SAMMANFATTNING PÅ SVENSKA ... 81
11 ACKNOWLEDGEMENTS ... 83
List of papers
1. Mahdi A, Hälleberg-Nyman M, Wretenberg P. Preoperative psy-chological distress no reason to delay total knee arthroplasty: a register-based prospective cohort study of 458 patients. Archives of Orthopaedic and Trauma Surgery 2020 1
2. Mahdi A, Hälleberg-Nyman M, Wretenberg P. Reduction in anxi-ety and depression symptoms one year after knee replacement: a register-based cohort study of 403 patients. Submitted.2
3. Mahdi A, Hälleberg Nyman M, Wretenberg P. How do orthopaedic surgeons inform their patients before knee
arthroplasty surgery? A cross-sectional study. BMC Musculoskelet Disord 2018, 19(1):414.3
4. Mahdi A, Svantesson M, Wretenberg P, Hälleberg-Nyman M. Patients' experiences of discontentment one year after total knee arthroplasty- a qualitative study. BMC Musculoskelet Disord 2020, 21(1):29.4
_______________________________________________________________________ 1 © 2020 Archives of Orthopaedic and Trauma Surgery
2 © 2020 European Journal of Orthopaedic Surgery & Traumatology (submitted)
Abbreviations and definitions
ANOVA Analysis of variance
ASA American Society of Anesthesiologists BMI Body mass index
CI Confidence interval
CSI Clinically significant improvement
EQ-5D-3L EuroQoL 5 dimensions instrument — 3-level version HADS Hospital Anxiety and Depression Scale
KOOS Knee Injury and Osteoarthritis Outcome Score LIA Local infiltration anesthesia
M Mean value
MRI Magnetic resonance imaging NPA No preoperative anxiety NPD No preoperative depression OA Osteoarthritis
p Probability value PA Preoperative anxiety PD Preoperative depression
PROM Patient-reported outcome measure SKAR Swedish Knee Arthroplasty Register SPSS Statistical Package for the Social Sciences TKA Total knee arthroplasty
TKR Total knee replacement VAS Visual analogue scale
1
Introduction
The main goal of this thesis was to increase the understanding of patients with co-existing knee arthritis and psychological distress before and after total knee replacement (TKR). An additional goal was to understand the opinion and experience of patients who are discontented after surgery, in the hope of improving patients’ contentment after TKR.
Discontentment can arise before, during, or after surgery, and has many possible reasons. Special consideration has been applied in the literature to patients with psychological distress (anxiety and/or depression symptoms). Studies show that this patient category might have a higher risk of discon-tentment after surgery, and so it is preferable to assess and manage these patients psychologically before surgery. However, this assessment should not mean excluding patients from surgery, as this might lead to paradoxi-cally increasing patient discontentment regardless of whether they undergo surgery or not. The present research project attempts to address some of these concerns.
2
Background
2.1
Osteoarthritis
Osteoarthritis (OA) is a common condition that can affect any joint in the body1-3. The most affected joints, however, are the knees, hips, hands, and spine. OA occurs when the cartilage in the joint becomes thinner, which in turn affects the smooth movements between the bones in the joint3-5. There are two types of OA: primary and secondary. It is still unclear what causes primary OA, but certain factors might initiate or promote its development. Age is one example. All people will eventually develop OA in one or more joints, and the relevant changes typically start to occur from the age of 40 onwards. However, not all people develop symptoms of OA. Gender is another factor, as OA affects women more often than men, and women usually have more severe symptoms. Genetic factors, occupation, and obesity are other factors that might predispose to primary OA1-7.
Secondary OA usually develops after an obvious insult to the cartilage1-3. Joint injuries, inflammatory joint diseases such as rheumatoid arthritis, and congenital joint conditions are all examples of factors that could pre-dispose to OA. The knee is one of the most common joints affected, be-cause the knees are both weight-bearing joints and more susceptible to loading and turning forces. As soon as arthrosis has developed in the knee, the body starts its healing mechanisms to protect the joint. The amount of synovial fluid increases, osteophytes develop, the subchondral bone starts to thicken, and loading in the other soft tissues changes. These compensa-tory body mechanisms are the basis of knee symptoms due to OA in the earlier and late stages3,7,8. Patients experience different severity of pain, swelling, stiffness, crepitations, locking, and changes in the joint align-ment. These symptoms affect patients’ daily activities, work, hobbies, and quality of life in different intensities 1-4,6,7,9.
Diagnosis of knee OA is usually settled with the aid of knee symptoms and radiology3,7. X-ray examination is an easy and common test to diagnose the severity of arthrosis (Figure 1). However, early OA changes might not be seen in conventional X-ray examinations. Magnetic resonance imaging (MRI) is more helpful in evaluating early cartilage changes. Knee arthros-copy is also helpful in evaluating the joint surfaces and cartilage status by
Laboratory blood tests and joint analyses are used to exclude other kinds of joint arthritis such as rheumatoid arthritis and septic arthritis1-4,6,7. Treatment of knee arthritis is indicated in the presence of symptoms1-3,6. The choice of treatment is based on the severity of symptoms, age of the patient, number of affected knee compartments, and time since the symp-toms started. Non-invasive treatment includes medication for pain man-agement, physiotherapy, and occupational therapy, while invasive proce-dures include injections of cortisone and lubrication, re-alignment osteot-omy, and joint replacement surgery1-4,6,7.
2.2
Total knee replacement
Total knee replacement (TKR), also known as total knee arthroplasty (TKA), is a surgical procedure in which the medial, lateral, and patello-femoral compartments of the knee joint are replaced with an artificial joint (Figure 2). The replacement is still considered total even in the ab-sence of a patellar button, because the femoral flange resurfaces half of the
compartment of the knee joint (medial, lateral, or patellofemoral) is re-placed with an artificial joint10-12. There are two general types of TKR: primary and revision replacement. Primary knee replacement means that the knee joint is replaced with an artificial joint for the first time (natural knee to artificial knee), while revision arthroplasty means that one or more components are replaced, added, or removed in a previously re-placed knee joint (artificial knee to artificial knee, artificial knee to ar-throdesis, or artificial knee to amputation)12,13.
TKR is a common surgical procedure throughout the world, and its fre-quency is increasing every year10,14. In Sweden, the number of TKR proce-dures increased from 70 to 150 per 100,000 inhabitants between 2000 and 201815. In the USA, TKR is expected to increase by 673% from 2005 to 203014. TKR is indicated mainly for a painful knee OA that affects the patient’s daily life. Fracture sequelae, osteonecrosis, and rheumatoid ar-thritis are other indications for surgery6,16-18.
Different surgical approaches are available for TKR12,19-21. The most com-mon approach is the standard medial parapatellar21. A vertical incision is made through the skin, the capsule is incised medially from the patella, the joint surfaces are exposed, cutting guides are used in sawing the joint sur-faces in the femur and tibia, the patellar joint surface is sometimes excised, and replacement of all the cut surfaces is performed. This step is often preceded by ligament balancing12. Other surgical approaches also exist, such as minimally invasive TKR and the mid-vastus and lateral parapatel-lar approaches12,22.
There are numerous prosthetic designs available, with some differences in the philosophy of design10,23. Generally, TKR prostheses are divided into cruciate-retaining and posterior stabilized designs. The first of these allows maintenance of the posterior cruciate ligament, while the other sacrifices it. TKR components are fixed to the bone surfaces with or without the aid of a synthetic material called cement, and so surgeries are classified into cemented and uncemented TKR10,12,23.
TKR results are generally promising in terms of alleviating knee symp-toms, improving knee function, and improving the patient’s quality of life16,24-28. Studies have shown levels of patient contentment varying from 70% to 92%. There are mechanical, psychological, and combined factors that could predispose to discontentment after this procedure13,16,18.
Previous studies have shown complication rates of 0.03–39% following TKR13,29. Complications can be classified into intraoperative (0.03–2.5%), early postoperative (0–39%), and late postoperative (0.3–27%)29. Exam-ples of intraoperative complications include neurovascular injury, extensor mechanism injury, and medial collateral ligament injury. Early postopera-tive complications occur within the first four weeks after surgery29,30, and examples include bleeding, superficial and deep infection, and thrombo-embolism. Examples of late postoperative complications include deep infection, aseptic loosening, instability, periprosthetic fracture, and prob-lems with the patella13,29,31. In Sweden, the risk has been estimated to be 3% for adverse surgical events, 1.7% for other medical events, 0.7% for cardiovascular events, and 0.13% for death within the first 90 days10. Infection is now the most common reason for revision (30%), followed by loosening (25%)10. Most complications are amenable to treatment. The high level of satisfaction after TKR despite the occurrence of complica-tions is likely due to its pain-relieving effect12,13,16,31,32.
There has been a continuous improvement of knee prosthetic design since Figure 2. X-ray following TKR of the right knee. Ref: Aamir Mahdi.
niques, with examples including patient-specific instrumentation, naviga-tion, and robotic surgery. All of these efforts were aimed at improving the precision of surgery, which could increase both patients’ contentment rate and prosthetic survival11,33,34.
TKR has proven to be a cost-effective surgical procedure in alleviating knee symptoms and improving function in patients with knee OA16,17,28,32,35-39. The mortality rate and risks of early complications are low after TKR40, but 6–30% of patients are still discontented after surgery regardless of the presence or absence of documented complications32,41-51. In Sweden, about 8% of patients with no documented complications re-main unsatisfied after TKR10,50.
Previous studies have investigated the relationship between patient discon-tentment and a diversity of mechanical, psychosocial, and combined fac-tors32,41-51. Discontentment is more common in patients with previous knee surgeries, higher American Society of Anesthesiologists (ASA) class, more co-morbidity factors, knee revision surgery for any reason, and complica-tions requiring hospital readmission32,41,48,50,51. These do not comprise all reasons for discontentment, as there are still discontented patients without documented complications, revision surgeries, or previous knee surgeries51,52. Moreover, there are also patients who consider their TKR to be successful despite complications, due to the resulting pain relief and improved function13,53.
Discontentment after TKR is also correlated with mechanical complica-tions and prosthesis malalignments. Rotational malalignment of prosthesis components could affect the patella tracking, which could lead to contin-ued pain after surgery54. Coronal malalignment has been shown to in-crease the rates of loosening and subsequent revisions55. There is an ongo-ing worldwide effort to improve the prosthesis design, surgical technique, and mechanical alignment in order to decrease the discontentment rate and increase prosthetic survival38,56,57. This was the basis for developing patient-specific instruments and robotic-assisted TKR to produce correctly aligned components22,33,34,38. However, there is some controversy in the literature about acceptable mechanical alignment. Many studies have shown that there are no significant differences in the clinical outcome and survivorship between aligned and mal-aligned TKR components36,58,59. Moreover, patients’ contentment one year after surgery does not correlate
with the mechanical axis or component alignment, but rather with the pain relief60-62.
In recent years, more focus has been placed on the effect of psychological factors in determining patients’ contentment after TKR41-43,46,63-65. Preoper-ative anxiety and depression have been considered as predictors of postop-erative patients’ contentment. One study showed higher levels of 90-day readmission, discontentment, and burden for health care providers among patients with depression66. Another study showed that patients with pre-operative anxiety and pain catastrophizing had more postpre-operative pain, poorer preoperative and postoperative knee function, and higher discon-tentment rates64.
These and other studies recommend preoperative psychological assessment and management to decrease the discontentment rate41-43,46,63-65. However, there is little knowledge about whether the symptoms of depression and anxiety are a result of symptoms and disabilities caused by knee OA. One study showed that increasing disability and poorer quality of life are rea-sons for higher anxiety prevalence among elderly people67. Others found that depression improved after hip/knee replacement surgery due to a re-duction of pain68, and that depression and anxiety symptoms improved 6 months after surgery69.
2.3
The Swedish Knee Arthroplasty Register
The Swedish Knee Arthroplasty Register (SKAR) was the first national arthroplasty register in the world, established in 197510,70,71. It has high completeness and correctness of data, helped by the simplicity of tracking patients through the unique personal identification number. Initially, SKAR was mainly focused on early complications and mortality. Howev-er, as knee replacement surgery proved to be a procedure with low mortal-ity and few early reported severe complications (except infection), the focus soon moved to the long-term outcome. Since the beginning, SKAR has worked to validate, reassess, and improve its data completeness and accuracy10,70,71.
Details recorded in SKAR include the diagnosis (OA, rheumatoid arthritis, others), type of surgery (TKR, unicompartmental knee arthroplasty,
revi-term complications, and patient satisfaction after surgery. This gives valu-able feedback regarding results of the surgery to clinics and surgeons, and facilitates continued improvements for the patients. SKAR data are also a valuable resource for scientific studies due to their high completeness, large number of cases, and long follow-up time. Information obtained from SKAR is population based and unselected, unlike information ob-tained from single centers and/or surgeons10,70,71.
SKAR started to evaluate data on patient-reported outcome measures (PROMs) in the 1990s. The SKAR PROM project started with one hospi-tal in 2018, and by 2019 24 out of 73 orthopedic departments were re-porting PROM data on a voluntary basis. Different knee-specific and ge-neric instruments have been used over the years. In the 1990s, SF-36, the Oxford Knee Score, and WOMAC were used, while today the PROMs consist of KOOS, EQ-5D-3L, EQ-VAS, VAS knee pain, and VAS satisfac-tion with surgery 10,70,71. Patients are requested to fill in these PROMs questionnaires 2–6 weeks before and one year after TKR.
Once the departments have entered their data into the relevant software program, SKAR analyzes and reports the anonymized data in its annual report and in real time on its publicly-available homepage (www.knee.se). Ethical approval is requested in the case of studies that need data from SKAR. Access can be provided only to data needed for a particular study10,70,71.
SKAR cooperates with other national arthroplasty registries in the Nordic countries via the Nordic Arthroplasty Register Association, as well as with international arthroplasty registers via the International Society of Arthro-plasty Registries, in order to standardize methods, common terms, and definitions70. It is currently in the process of merging with the Swedish Hip Arthroplasty Register to become the Swedish Joint Arthroplasty Register.
2.4
Patient-reported outcome measures
PROM questionnaires are used to measure outcomes from the patient’s perspective72,73, and can be used both before and after the treatment. They are classified into generic and disease-specific instruments; generic ments measure the general health outcome, while disease-specific instru-ments measure outcomes specific to the disease. The PROMs used in this
thesis have all been psychometrically validated72,73, and are described in more detail below.
2.4.1
Knee Injury and Osteoarthritis Outcome Score
The Knee Injury and Osteoarthritis Outcome Score (KOOS) is a disease-specific PROM developed to assess patients’ experiences of their knee and associated problems. It can be used to evaluate both short-term and long-term consequences of knee injury, osteoarthritis, and knee surgery74. KOOS consists of 42 items divided into 5 subscales: pain (9 items), other symptoms (7 items), activities of daily living (17 items), sport and recrea-tion funcrecrea-tion (5 items), and knee-related quality of life (4 items). It is re-garded as a feasible instrument that usually takes 10 minutes to fill in. However, it must be remembered that 10 minutes can be a long time when patients are asked to fill in multiple questionnaires simultaneously and in a frenetic clinical environment74.
KOOS has been validated in different populations and age groups, for different knee complaints, and after both non-surgical and surgical treat-ments. Relevance and comprehensiveness were evaluated by asking pa-tients about their degree of understanding of the construct and judging the importance of the items (content validity). Over 90% of those asked re-ported improvement in all the four subscales; pain, symptoms, activities of daily living, and knee-related quality of life were all important when de-ciding to have their knees operated on. The sport/recreation subscale was also reported as being important, but only by 51% of patients74-76.
Experts consider KOOS to have good face validity. In terms of construct validity, there is a strong correlation between KOOS and the SF-36 physi-cal function subsphysi-cale, a strong correlation between KOOS and the EQ-5D questionnaire, and a weak correlation between KOOS and SF-36 mental function. This reflects the fact that KOOS was designed to measure physi-cal rather than mental or emotional properties 74,75,77,78.
Reliability was validated by testing internal consistency and test-retest reliability. Internal consistency is the degree of interrelatedness among the items. KOOS was shown to have a good internal consistency (Cronbach’s α > 0.70) for all individual subscales. A similar trend was shown for
test-retest reliability for all the subscales (intraclass correlation coefficient > 0.70)74,75,78,79.
KOOS has shown good sensitivity in different knee conditions (OA, me-niscal injury, cartilage damage, and cruciate ligament injuries), different treatment modalities (surgical and non-surgical treatment), and different surgical management (cruciate ligament reconstruction, cartilage repair, TKR). It has also been used in non-knee related conditions, specifically the foot and hip74,75,77,78.
Responsiveness is defined as the ability of a measure to detect a meaning-ful clinical change over a time. KOOS has a good responsiveness, particu-larly in the subscales for knee-related quality of life and pain. The sport/recreation subscale has the worst responsiveness. Effect size is one way to study responsiveness. Previous studies have shown that surgical treatment has a larger effect than non-surgical treatment, particularly in the case of TKR, which has the largest reported effect size74,75,78,79.
KOOS is a self-reported instrument, and does not reflect the magnitude of performance deficit after surgery. The literature suggests that it is best to combine KOOS with subjective performance assessments in order to get the best outcome measures. However, KOOS is regarded as being psy-chometrically well analyzed74,75,78.
2.4.2
EQ-5D-3L
The EuroQoL five-dimension three-level scale (EQ-5D-3L) was designed by the EuroQoL group in 199080. It is a generic (general) self-report in-strument81 comprising of five subscales: mobility, self-care, usual activities, pain/discomfort, and anxiety/depression. Each subscale has three levels: no problems, some problems, and extreme problems. It takes less than one minute to complete, and is available in both paper and digital form. It has been translated into 200 languages80,81.
EQ-5D-3L has adequate validity, very good feasibility, good responsive-ness in certain conditions, and moderate reliability81-86. However, the ex-istence of a ceiling effect means that EQ-5D-5L might be preferable87,88. The weak sensitivity of EQ-5D-3L makes it less useful in cost-effect analy-sis, where it is instead recommended to use a disease-specific instrument 80-86,89.
EQ-5D-3L is included in the SKAR PROM project10. Patients fill in the questionnaire 2–6 weeks prior to TKR and one year after. There are three main types of change: worse (no problems before and some problems af-ter; no problems before and extreme problems afaf-ter; some problems be-fore and extreme problems after), unchanged (no problems both bebe-fore and after; some problems both before and after; extreme problems both before and after), and better (extreme problems before and some problems after; extreme problems before and no problems after; some problems before and no problems after). The 2019 SKAR annual report showed that just over half of patients experienced improvements in their mobility and pain relief, one third improved in their usual activities, one fifth improved in their anxiety/depression, and few improved in self-care. Similar propor-tions of improved patients were shown in previous years10,15.
The EQ visual analogue scale (EQ-VAS) was also designed by EuroQoL group81, and is usually used in parallel with EQ-5D-3L. EQ-VAS consists of a scale running from 0 to 100 that is used to measure self-reported gen-eral health. Zero represents the worst health while 100 represents the best health. This is also used by SKAR together with data on other PROMs15,81.
2.4.3
Visual analogue scale
The visual analogue scale (VAS) is a self-reported instrument which is used for measuring both acute and chronic pain90. This has a particular im-portance in evaluating the effect of pain treatment such as pain medica-tion, physiotherapy, occupational therapy, or surgical treatment. It has been widely used since 1920 in many epidemiological and clinical studies90-92. The scale ranges from 0 to 100, where 0 represents no pain and 100 represents the most severe pain that the patient can imagine. It is available in two forms: paper and digital90,93,94.
VAS psychometrics have been studied previously93-96. It is a feasible in-strument which takes less than one minute to complete. It is more valuable in measuring changes within individuals, and less valuable in comparing groups at one time point. Its simplicity means that it has high acceptability among patients. It shows good reliability, particularly among literate peo-ple, and its validity is also good. VAS has a high responsiveness through its ability to detect clinically significant changes, though the minimum
VAS is commonly used to evaluate the intensity of pain experienced after TKR. Several arthroplasty registries and previous studies have used VAS because of its simplicity as well as the fact that pain is regarded as the main indication for TKR 10,97. SKAR uses VAS knee pain, with a difference of 15–20 points considered clinically relevant. The most recent SKAR annual report showed a mean difference of 47 points between preopera-tive and one-year postoperapreopera-tive VAS pain for the participating hospitals10,98.
2.4.4
Hospital Anxiety and Depression Scale
The Hospital Anxiety and Depression Scale (HADS) is a self-reported instrument that was designed in 1983 to determine the level of anxiety and depression in patients suffering from physical illness99. It is not a diagnos-tic tool, but rather a screening tool. Many diseases and medical conditions are associated with varying degrees of emotional distress both before and during treatment100. Exaggeration of the symptoms might misleadingly indicate that the disease is severe. On the other hand, if the patient is hid-ing their anxiety or depression symptoms, this might affect the decision to treat the condition. For instance, a surgeon might decide to postpone the surgery because the patient might show his/her symptoms with a milder reaction than the symptoms bothering. Therefore, HADS was designed mainly to explore the symptoms of anxiety and depression100-102.
HADS consists of 14 items; 7 for anxiety and 7 for depression99,100,103. Each item has four possible answers (0 to 3), and so total scores range from 0 to 21 points for both anxiety and depression. A review of a large number of studies identified a cut-off point of 8/21 for anxiety or depres-sion, meaning that people who score less than 8 are considered not to have anxiety/depression symptoms100-102.
HADS takes 2–5 minutes to complete99-103, and has been validated in sev-eral languages. Although it was initially designed to be used in the hospi-tal, studies have shown its usefulness in primary care and the general pop-ulation. It has good validity in defining anxiety and depression and as-sessing their severity in both somatic and psychotic cases, and is sensitive to change both during the disease course and in response to intervention. HADS has been used as a predictive tool for both psychosocial and physi-cal outcomes, and has also been validated in adolescent patients. It has
good psychometric properties, and can discriminate between adolescents with anxiety or depression and those without these symptoms100-102,104. One limitation of HADS is that illiterate people can find it difficult to understand the questions and hence to answer correctly as they intrinsical-ly feel100. Moreover, some people feel ashamed to show their psychological distress, and do not answer in a way that reflects their true feelings. An-other weakness is a substantial ceiling effect in items 7 and 8100-102.
Only a few knee surgeons in Sweden routinely use HADS as a screening tool before assigning patients for TKR105. Preoperative screening with HADS might be of great value in assessing patients’ preoperative psycho-logical distress.
2.5
Contentment after TKR
Contentment after TKR is defined as a patient being satisfied with the outcome of knee replacement surgery. It is often related to the fulfillment of the patient’s pre-surgical expectations32,39,44-46,48,106. These expectations vary among patients; some expect their knee pain to improve enough that they can perform their daily activities with only mild discomfort, while others expect to get a new joint with similar function to a natural knee. There is a variety of nuance between these two extremes. Discontentment after TKR does not necessarily mean that the patient is dissatisfied; some patients experience discontentment because they have gained some or many of the expected results, but not all107,108. For instance, a patient who expected to be pain-free and retain the normal knee range of motion after TKR may be discontented despite becoming pain free if their knee range of motion was acceptable but not to the extent that they had expected. SKAR categorizes patient satisfaction into five levels (very satisfied, satisfied, moderate satisfied, non-satisfied and very unsatisfied) on the basis of pa-tients’ postoperative satisfaction score10.
I believe that the word “discontentment” is a better way to describe pa-tients’ nuance of non-satisfaction, not only in this thesis but also in other studies. However, in this thesis the words “satisfaction”, “non-satisfaction”, and “dissatisfaction” are sometimes used because they have been mentioned previously in the literature.
2.6
Rationale
Several studies have considered preoperative psychological distress as a predictor of postoperative discontentment after TKR, but few have fo-cused on the actual benefits of TKR in patients with preoperative psycho-logical distress.
It is uncertain whether preoperative psychological anxiety and depression symptoms are secondary to knee symptoms and the subsequent disabili-ties, or if there are other causes. Analysis of this issue might give a better understanding and improve assessment of preoperative psychological symptoms.
Orthopedic surgeons play an essential role in patient contentment after TKR, not only because of their skill in performing surgery, but also be-cause of the effect of their attitude in informing patients, assessing psycho-logical distress, and selecting patients who are appropriate for such sur-gery. Little knowledge is available about orthopedic surgeons’ attitudes to patients who will undergo TKR.
Most of the previous studies on patients’ contentment after TKR had a quantitative design, and to my knowledge there have been no qualitative studies focusing on patients’ experiences of discontentment after surgery. Conducting such a study might be helpful in analyzing information that could not be captured by quantitatively designed studies, and could illu-minate some facts about the controversies in previous studies.
3
Aims of the thesis
The overall aim of this thesis was to study patients with anxiety and/or depression symptoms before and after primary total knee replacement, particularly in terms of improvements in patient-reported outcomes, and particularly among those who were discontented with the surgery. This could hopefully improve patients’ contentment after TKR.
Specific aims and objectives were:
Study I
To compare PROMs before and after total knee replacement between patients with preoperative anxiety and/or depression symptoms and those without such symptoms.
Study II
To investigate the change in the prevalence of anxiety and depression symptoms one year after primary total knee replacement, and to compare demographic data between patients with and without anxiety and depres-sion symptoms preoperatively and postoperatively.
Study III
To investigate how TKR surgeons in Sweden inform their patients before surgery.
Study IV
To capture patients’ experiences of discontentment one year after primary TKR without documented complications.
4
Patients and methods
Different study design, methods, and analyses were used based on the aims of the studies. Table 1 summarizes the patients and methodology in the four studies.
Table 1. Overview of the studies in this thesis.
Study I II III IV Design Prospective cohort Prospective cohort Cross-sectional (descriptive) Qualitative Participants n=458 TKR patients n=403 TKR patients n=219 Knee replacement surgeons n=44 TKR patients
Data collection KOOS, VAS,
5D, EQ-VAS, HADS
KOOS, VAS, EQ-5D, EQ-VAS, HADS Study-specific questionnaire Individual interviews
Data analysis Paired t-test, χ2
test, independ-ent t-test, two-way ANOVA χ2 test, independent t-test, McNemar’s test, η2 value Descriptive data analysis, χ2 test, thematic analysis Inductive qualitative content analysis
4.1
Settings
The orthopedic departments in Örebro County, Sweden, comprised the main setting of Studies I, II, and IV. At the beginning of the studies and data collection, there were three orthopedic departments located in one university hospital and two regional hospitals: Örebro, Karlskoga, and Lindesberg. All three were public hospitals. In 2017, these three orthope-dic departments merged into one. One hospital became the main center for reconstruction surgery including TKR, while the other two hospitals grad-ually decreased the number of arthroplasty surgeries. The orthopedic de-partment covers an area of 300,580 inhabitants. Patients with knee OA or
other diagnoses are scheduled for surgery which should be performed within 90 days according to the Swedish health care guarantee.
For Study III, a study-specific questionnaire was distributed in paper form to all 65 TKR departments in Sweden in May 2016. A list of all orthope-dic departments performing TKR was obtained from SKAR.
4.2
Samples
Study I
Study I included a consecutive sample of patients who underwent TKR between April 2016 and July 2018. The inclusion criteria were having knee arthritis (primary OA, secondary OA, and inflammatory arthritis) and being scheduled for primary TKR. The exclusion criteria were revision arthroplasty, unicompartmental arthroplasty, and patella-femoral arthro-plasty.
During the study period, 719 patients underwent knee arthroplasty. The participation response rate was 91%. After excluding patients who de-clined (n=65), those with incomplete data (n=157), and those who did not meet the inclusion criteria (n=39), 458 patients with complete data were included in the study (Figure 3).
Study II
Study II included a consecutive sample of patients scheduled for knee re-placement surgery between April 2016 and April 2018. The inclusion criteria were having knee OA and being scheduled for primary knee re-placement surgery at one of the three hospitals. The exclusion criteria were being scheduled for a revision of knee arthroplasty or unicompart-mental knee arthroplasty, or having a diagnosis other than OA. Of the 610 patients who underwent surgery between April 2016 and April 2018, 207 were excluded due to declining participation (n=37), unicompart-mental knee arthroplasty (n=22), revision surgery (n=12), diagnosis other than OA (n=22), and incomplete data (n=114), leaving 403 included in the study (94% response rate) (Figure 4).
Study III
Of the approximately 311 knee arthroplasty surgeons in Sweden, 219 answered the questionnaire. Thus, about 70% of the surgeons who regu-larly perform TKR surgery in Sweden answered the questionnaire. Accord-ing to the orthopedic chief managers, the reasons for non-response were that the surgeons were unwilling or not interested. Non-response was not related to the kind of hospital (university hospital, county council hospital, or regional hospital).
Study IV
Of the 348 patients who underwent TKR in 2015–2016 at the three hospi-tals and who responded to our letter, 61 patients (18%) reported discon-tentment. Of these 61 patients, 7 were excluded due to documented post-operative complications, and another 10 declined to participate due to logistical reasons such as difficulty coming to the interview, illness, or late response. The final sample consisted of 44 patients (Figure 5). Of these, 31 had undergone TKR surgery for the first time while the remaining 13 had undergone TKR previously in the other knee.
Figure 3. Flow chart of patient inclusion in Study I. 654 patients (91% response rate)
497 patients
458 patients included
65 declined to participate
39 did not fulfill the inclusion criteria
157 returned incomplete pre- and/or postoperative questionnaires 719 patients underwent TKR,
Figure 4. Flow chart of patient inclusion in Study II. 610 patients underwent TKR, 2016–2018 573 patients 517 patients 403 patients included 37 declined to participate
56 did not fulfill the inclusion criteria
114 returned incomplete pre- and/or postoperative questionnaire
Figure 5. Flow chart of patient inclusion in Study IV. 348 patients (97.7%) responded 7 (2%) non response 1 (0.3%) dead 61 patients (17.2%) were discontented 287 (80.6%) contented 54 patients (15.2%) had no complications 7 (2%) with documented complications 44 patients (12.4%) included 10 (2.8%) declined participation 356 patients sent letters one
4.3
Procedures
Studies I & II
All patients scheduled for TKR were informed about the study. They were asked to complete a questionnaire including PROMs, regardless of wheth-er they wwheth-ere included in the study. PROM data collection before and one year after TKR is routinely performed by the departments where the study was conducted, and these data are then reported to SKAR. We added the HADS questionnaire to the other PROMs.
Study I was expected to include 500 patients during the study period (2016–2018), which was considered a representative sample if the TKR population was around 750 patients109. It was decided to include the first 500 consecutive patients with complete data.
For Study II, we expected that 10–20% of patients (n=30–60) would have symptoms of anxiety and/or depression, and hence estimated that enrol-ling about 300 patients would be enough to produce statistically signifi-cant differences with a power of 0.8 and alpha of 0.05. However, we en-rolled 403 patients to compensate for eventual attrition. SKAR provided us with the PROM questionnaires used by the registry, provided the soft-ware for entering and calculation of the HADS score, and gave us access to the data for patients operated in the three Örebro County hospitals. Thus, we took advantage of SKAR routines in terms of using a validated questionnaire and a software program.
Despite an organized study protocol, missing data were inevitable. Rea-sons for missing data in Studies I and II were:
1. At the beginning of the study, the nurses had not yet got into the routine of distributing the questionnaire, which caused missing patients
2. Patients delaying their answers for more than 3 months 3. Incomplete answering of the questionnaire
4. Patients not returning the preoperative or postoperative question-naire despite reminders
Study III
A study-specific questionnaire was distributed in paper form to all 65 TKR departments in Sweden in May 2016. A reminder letter was sent in October 2016 to increase the response rate. For each department, the questionnaire was sent to the SKAR contact surgeon at each hospital, who then distributed the questionnaire to all the orthopedic surgeons who per-formed TKR at that department. Arthroplasty departments in Sweden are located in either publicly-funded university hospitals, regional hospitals, county hospitals, or private clinics.
The initial response rate was 55%, but the reminder letter increased this to 92% (60 departments). The five orthopedic departments that failed to return the questionnaire were located in one university hospital and four county hospitals.
Study IV
All 356 patients who had undergone primary TKR in 2015–2016 at the three hospitals were sent a letter one year after surgery, asking whether they were “satisfied” or “dissatisfied” with the TKR. Dissatisfied patients with no documented complications were invited to participate in the inter-view study. The author of this thesis (AM), who is an orthopedic surgeon, conducted all except one of the 44 interviews. The last author of Study IV (MHN) conducted one interview because the participant was the present author’s patient. The audio-recorded interviews varied between 15 and 57 minutes (mean 35 minutes) and were transcribed verbatim by a profes-sional secretary.
4.4
Outcome measures
Studies I & II
The data for Studies I and II were collected via HADS and four PROM questionnaires: KOOS, EQ-VAS (general health), VAS pain, and EQ-5D-3L. All these questionnaires were filled in by the patients at two time points: before the surgery and one year after.
Preoperative VAS expectations were assessed on a 0–100 scale, with 0 representing no pain and 100 the worst possible pain. Postoperative VAS
pleted by the patients one year after TKA and indicated the degree of satis-faction with their knee pain. Another 0–100 scale (best–worst) along with the question “How good is your knee?” was used to assess postoperative patient satisfaction after one year15.
Study III
A study-specific 16-item questionnaire was used to investigate how infor-mation was given to and discussed with patients before surgery. The first 14 questions covered written and oral information, patients’ expectations, surgery on mild indication, and patients’ psychiatric history. The choice of questions was based on evidence from the literature that these factors can have an impact on outcome48,110-120. A five-point Likert scale was used, with response alternatives of always, often, sometimes, seldom, and never; this was chosen after discussion with a statistician (Table 2). The final two questions considered what the surgeons believed were the most common reasons for dissatisfaction after TKR, and were answered in the form of free text. There was no previous validated questionnaire covering all the items that we intended to study. The questionnaire was written in Swedish by the author of this thesis and his main supervisor. The co-supervisor made then several changes to the language and the construct. The ques-tionnaire was then sent to two independent orthopedic surgeons for com-ments, and after a few more changes the questionnaire was finalized.
Table 2. The questions answered by knee surgeons in Study III.
No. Question text
1 Do you use written information with clear text about the risks, complications, and benefits of TKA surgery?
2 Do you use oral information with clear text about the risks, complications, and bene-fits of TKA surgery?
3 How many years have you worked with knee arthroplasty? 4 How many primary TKAs do you perform per year?
5 How many knee revision arthroplasties do you perform per year?
6 What is your percentage of working with knee replacement surgery? a. 100 % knee specialist, b. 75% c. 50% d. 25%
7 Do you ask your patients about their expectations?
8 Do you inform your patients that about 20% of patients are not satisfied after TKA surgery despite the absence of obvious explanation?
9 If you see a patient with severe knee pain, mild radiological arthrosis, a great desire to be operated, and anxiety/depression, how often do you proceed with surgery? 10 Do you routinely enquire about psychiatric history?
11 If you realize that a patient has depression/anxiety, do you consult a psychiatrist before proceeding with surgery?
12 Do you use a psychiatric enquiry sheet for evaluation of psychiatric problems? 13 If you use a psychiatric enquiry which reveals depression or anxiety, do you make sure
that the patient receives psychiatric treatment before proceeding to knee surgery? 14 Do you think that psychiatric problems play some role in the results of knee surgery? 15 Many dissatisfied patients describe pain. Aside from radiating pain, what do you think
is generally the most important reason for pain?
16 What do you think is the single most important reason for a patient to be dissatisfied after TKA?
Study IV
Semi-structured face-to-face interviews were conducted one year after knee replacement surgery at locations chosen by the participants; either in the hospital (n=41) or at their homes (n=3). The two main questions were “Can you tell me about the time before surgery?” and “Can you tell me about the time after surgery?” Follow-up questions such as “Please tell me more about that,” or “Can you give an example?” were used. The inter-views also included questions regarding the information given before and after surgery, patients’ expectations, and improvements fields after sur-gery. The interview guide was developed specifically for this study (Table 3). Regret of surgery was also investigated, as this might indicate discon-tentment regarding TKR surgery.
Table 3. Interview guide: English version.
1.
“Can you tell me about the time before surgery?”• Follow-up questions: “Please tell me more about that,” “Can you give an example?”, “What do you mean?”, “Would you like to explain more?”, “Can you please elaborate more on that?”, “What did you think?”, and “Explain more?”
• Please tell me about your expectations for the knee surgery.
• In what way and to what extent have these expectations been met or not been met?
2.
“Can you tell me about the time after surgery?”• Follow-up questions: “Please tell me more about that,” “Can you give an example?”, “What do you mean?”, “Would you like to explain more?”, “Can you please elaborate more on that?”, “What did you think?”, “Explain more?”
• Do you regret the knee surgery? Please tell me why/why not.
• If you could ask for anything at all, what improvements do you suggest would have optimized your experience and the results of your knee re-placement surgery?
4.5
Demographic variables
Study I
Demographic variables collected in Study I were gender, age, BMI, ASA classification (I, II and III), Charnley class (A, B1, B2, and C), diagnosis (OA, osteonecrosis, rheumatoid arthritis, and fracture sequelae), Outcome Measures in Rheumatology and Osteoarthritis Research Society Interna-tional responder criteria (OMERACT-OARSI-responder), with or without patella component, previous knee surgery, type of anesthesia (general, spinal), local infiltration anesthesia (LIA), use of tourniquet during sur-gery, surgical time, and side being operated (right or left knee). Demo-graphic variables collected for patients with missing data were gender, age, BMI, ASA class, Charnley class, preoperative anxiety symptoms, and pre-operative depression symptoms where available.
Charnley class is used routinely by SKAR10, and comprises three catego-ries. Class A patients have unilateral knee disease. Class B patients have with bilateral knee disease, and are further subdivided into Class B1 when the contralateral knee is not resurfaced and Class B2 when the contrala-teral knee is resurfaced. Finally, Class C patients have bilacontrala-teral knee dis-ease along with multiple joint complaints or other co-morbidity which affects walking ability.
OMERACT-OARSI-responder is also used routinely by SKAR10. It ana-lyzes the change in WOMAC (Western Ontario and McMaster Osteoar-thritis Index) pain, function, and total score one year after TKR, and cate-gorizes patients into “responders” and “non-responders”. A responder is a patient who has improved by 20–50% and at least 10 points in two of WOMAC pain, function, and total score. A non-responder is a patient who does not meet the criteria to be a responder.
Study II
Demographic variables collected in Study II were age, gender, BMI, ASA class, surgical time, VAS pain, EQ-5D index, EQ-VAS, preoperative VAS expectation, “How good is your knee?” after one year, KOOS for symp-toms, pain, ADL function, sport/recreation function, and knee-related quality of life.
VAS expectation is used routinely by SKAR10. It comprises a scale from 0 to 100 where 0 represents no pain and 100 represents extremely severe pain. Patients are asked preoperatively to score their expected VAS one year after surgery.
“How good is your knee?” is also used routinely by SKAR10. This also comprises a scale from 0 to 100, and here 0 represents a very good knee and 100 represents a very bad knee. Patients are asked one year after TKR to score their knee based on their own perception of improvement. Demographic variables collected for patients with missing data were gen-der, age, BMI, ASA class, Charnley class, preoperative anxiety symptoms, and preoperative depression symptoms.
Study III
The surgeons were divided into subgroups classified in terms of years of experience, the volume of primary TKR/revision surgeries per year, and the percentage of work done on knees.
Study IV
Demographic variables collected for study IV were gender, age, BMI, ASA class, background (Swedish vs. non-Swedish), hospital (1, 2, or 3), type of prosthesis (Genesis II, NexGen cemented, NexGen non-cemented, and Journey), and side (right or left knee).
5
Analysis
5.1
Study I
Patients were first categorized based on HADS score into two groups: one with anxiety and another without anxiety. Similarly, two groups were obtained with and without depression symptoms. Patients were regarded as having anxiety or depression symptoms if their relevant HADS score was more than seven99.
A χ2 test was used to compare binary data, while an independent t-test was used to compare continuous data on demographic variables. A paired t-test was used to compare changes in mean differences in each group before surgery and one year after surgery regarding the outcome measures (KOOS subscales, VAS pain, EQ-VAS, and EQ-5D index). Patients were divided into four groups (with preoperative anxiety symptoms, without preoperative anxiety symptoms, with preoperative depression symptoms, and without preoperative depression symptoms), and a paired Student’s t-test was used to compare the mean differences in outcome measures in all four groups before and one year after TKR. Two-way repeated measures ANOVA was used to compare the mean differences within each group (anxiety, no anxiety, depression, and no depression) and the mean differ-ences between groups (anxiety symptoms vs. no anxiety symptoms and depression symptoms vs. no depression symptoms).
A χ2 test was used to compare patients with and without preoperative anxiety symptoms in relation to clinical improvement in EQ-5D index, VAS pain, EQ-VAS, and the KOOS subscales. A p-value of less than 0.05 was regarded as statistically significant. Specific cut-off points were used to determine whether an improvement of an outcome measure was clini-cally significant (CSI): 0.25 for 5D index, 15 for VAS pain and EQ-VAS, and 8 for the KOOS subscales 10,80,121.
Similar comparisons were performed for patients with and without pre-operative depression symptoms. Thus, cross tables were used to compare two categories of preoperative psychological state (with anxiety/depression and without anxiety/depression) in relation to two categories of 1-year postoperative outcome measures (clinically significant improvement and
Version 25 of SPSS (IBM, SPSS Inc., Chicago, Illinois, USA) was used for the data analysis.
VAS expectations were sorted into three categories: high expectations (0– 30), moderate expectations (30–60), and low expectations (60–100). Simi-lar categories were formed for VAS satisfaction (i.e., high/very high satis-faction, moderate satissatis-faction, and low/very low satisfaction). The “How good is your knee?” variable was categorized as good/very good, moder-ately good, and bad/very bad. Mean values for VAS expectations, VAS satisfaction, and “How good is your knee?” were also calculated.
5.2
Study II
Patients were first categorized based on HADS score into two groups: one with anxiety symptoms and another without anxiety symptoms. Similarly, two groups were obtained with and without depression symptoms. Pa-tients were regarded as having anxiety or depression symptoms if their relevant HADS score was more than seven99.
A χ2 test was used to compare the two groups with/without anxiety before and one year after surgery. The dependent variable was anxiety symptoms pre-surgery, the independent variable was TKR, and the control variable was anxiety symptoms one year postoperatively. An independent t-test was used to compare demographic data between the two groups with and without anxiety symptoms before surgery. Similar statistical tests were performed for depression symptoms.
McNemar’s test was used for pre- and postoperative comparisons of anxi-ety and depression symptoms. Odds ratios (ORs) with 95% confidence intervals (CIs) were calculated when appropriate. Differences were regard-ed as statistically significant when the p-value was less than 0.05.
The effect size was calculated by measuring the η2 value using SPSS with both cross-table comparison and univariate analysis of variance. Small, medium, and large effect sizes were defined as 0.01, 0.06, and 0.14 respec-tively, and a large effect size was considered to be clinically important. Version 25 of the SPSS software package (IBM, SPSS Inc., Chicago, Illi-nois, USA) was used for the data analyses.
5.3
Study III
The statistical data analysis was divided into two parts:
General analysis: The data were described in terms of frequencies of an-swers. Duration of experience was not taken into consideration.
Specific analysis: The surgeons were divided into subgroups classified in terms of years of experience, volume of primary TKR/revision surgeries per year, and percentage of work done on knees. For statistical purposes, orthopedic surgeons who performed 75% of their work on knees were grouped with those working solely on knees. Subgroup analysis was per-formed to see if there were differences between the groups.
Data were analyzed as frequencies of each answer (always, often, some-times, seldom, and never). More specifically, subgroup data analysis was performed to see if there were any statistically significant differences. The χ2 test was used to compare multiple categorical groups. Version 24 of the SPSS software package (IBM, SPSS Inc., Chicago, Illinois, USA) was used for the quantitative data analysis.
The two open questions were analyzed qualitatively with a quantitative component. The text was coded inductively and grouped into five main categories based on 262 responses, and then analyzed with a method based on thematic qualitative analysis122,123. Initial coding was performed by the author of this thesis (AM). Next, the second author of the study (MH) checked the quality of the coding and changes were made until con-sensus was reached. To facilitate thematic analysis, responses on the two free-text items were uploaded into version 11 of the NVivo software package (Boston, MA, USA).
5.4
Study IV
The interviews and transcripts were analyzed by means of qualitative con-tent analysis with an inductive approach124,125. Version 11 of the NVivo software package (Boston, MA, USA) was used to facilitate the categoriza-tion. Each interview was listened to while reading the text, and the tran-scripts were read meticulously. Meaning units that related to the aim of the study were extracted and condensed, and then a code was created for
for similarities and variations and sorted into categories. The sub-categories were grouped into generic sub-categories, and finally a main catego-ry emerged. Data analysis was initiated by the author of this thesis (AM), but the progress of each step in the analysis was scrutinized and discussed between AM and the last author of the study (MHN). The analysis was co-assessed by the second author (MS) and then iteratively revised be-tween these three authors until final agreement was established.
6
Ethical considerations
All the studies were carried out in accordance with the Helsinki Declara-tion and approved by the Regional Ethical Review Board of Uppsala, Sweden (refs: 2016/191 for Studies I and III, and 2016/191 [2019-02077] for Studies II and IV).
No author had any conflict of interest. All participants gave their verbal and written consent before surgery, and a new written consent form was sent together with the questionnaire one year after surgery (Studies I and II).
The data collected for Study III did not include the identity of the surgeon. However, written information and consent were sent together with the questionnaire to every participant. Participants’ names and personal in-formation were not requested.
The participants in Study IV gave their written informed consent, and were informed that participation was voluntary and that they could with-draw from the study at any time. New verbal information was given be-fore each interview.
Data for all the studies were treated with full confidentiality according to Swedish law. Data entry and analysis were carried out with a secure soft-ware program provided by SKAR. Access to SKAR data was restricted to the study data. No sensitive personal information was explored in the studies.
7
Summary of results
7.1
Study I
The data were first categorized according to the preoperative anxiety and depression state: 70 patients (15.3%) had preoperative anxiety symptoms and the remaining 388 (84.7%) did not, while 44 patients (9.6%) had preoperative depression symptoms and the remaining 414 (90.4%) did not.
Analysis of demographic data (Table 4) showed no statistically significant differences between the groups with and without preoperative anxie-ty/depression, except that patients with preoperative anxiety/depression were younger than patients without anxiety/depression (68 vs.70, p=0.01 and 66 vs. 70, p=0.003, respectively), the group with preoperative depres-sion symptoms had higher BMI than the group without preoperative de-pression symptoms (31 vs. 29, p=0.001), and the proportion of women was higher in the group with depression symptoms than in the group without (66% vs. 51%, p=0.05).
Comparisons were performed for each group before surgery and one year after surgery. All groups showed a statistically and clinically significant improvement in all the outcome measures (Tables 5 and 6).
Patients with preoperative anxiety and/or depression generally showed less improvement in the outcome measures compared to the groups without anxiety/depression, but most of the differences were not statistically signif-icant except in three variables. First, the mean difference in KOOS sport/recreation function was significantly higher in the non-anxiety/non-depression groups than in the anxiety/non-anxiety/non-depression groups (Tables 5 and 6). Second, a clinically significant improvement in EQ-VAS was seen in 55% of patients with preoperative anxiety versus 38% of patients without pre-operative anxiety (p=0.004), and in 52% of patients with prepre-operative depression versus 38% of patients without preoperative depression (p=0.04) (Tables 7 and 8). Third, clinically significant improvements in KOOS-pain were seen in 95% of patients without preoperative anxiety versus 89% of patients with preoperative anxiety (p=0.03) (Tables 7 and 8).
Our data indicated that 92% of patients had high preoperative VAS ex-pectations, 81% had high/very high VAS satisfaction, and 70% had good/very good knee one year after TKR. Patients without preoperative anxiety and/or depression had slightly higher values for the preoperative VAS expectations, VAS satisfaction, and “How good is your knee?” vari-ables. However, the differences in proportions and mean values were not statistically significant except for the “How good is your knee?” variable, where the mean value was statistically significantly higher in the non-anxiety group than in the non-anxiety group (23 vs. 29, p=0.03) (Table 9).
Table 4. Demographics and other comparative variables for the patients in Study I (n=458).
Variable PA (n=70) NPA (n=388) p-value PD (n=44) NPD (n=414) p-value* Sex, n (%) Female 43 (61) 196 (51) 0.09 29 (66) 210 (51) 0.05 Age, m (SD) 68 (8) 70 (8) 0.01 66 (9) 70 (8) 0.003 BMI, m (SD) 30 (4) 29 (4) 0.08 31 (4) 29 (4) 0.001 ASA class, n (%) I II III 18 (26) 45 (64) 7 (10) 99 (25) 255 (66) 34 (9) 0.94 15 (34) 24 (55) 5 (11) 102 (25) 276 (66) 36 (9) 0.27 Charnley class, n (%) A B1 B2 C 14 (20) 16 (23) 7 (10) 32 (46) 98 (25) 97 (25) 58 (15) 131 (34) 0.37 7 (16) 8 (18) 6 (14) 22 (50) 105 (25) 105 (25) 59 (14) 141 (34) 0.21 Diagnosis, n (%) Osteoarthrosis Osteonecrosis Rheumatoid arthritis Fracture sequelae 65 (93) 2 (3) 2 (3) 1 (1) 377 (97) 5 (1) 3 (1) 3 (1) 0.29 42 (96) 1 (2) 0 (0) 1 (2) 400 (97) 6 (1) 5 (1) 3 (1) 0.61 OMERACT-OARSI-responder, n (%) Non-responder Responder 12 (17) 58 (83) 38 (10) 346 (90) 0.15 4 (9) 40 (91) 46 (11) 364 (89) 0.39 Patella component 0 (0) 9 (2) 0.19 1 (2) 8 (2) 0.87 Previous surgery, n (%) 13 (19) 86 (22) 0.34 10 (23) 89 (20) 0.63 Anesthesia, n (%) General Spinal 40 (57) 30 (43) 221 (57) 167 (43) 0.97 21 (48) 23 (52) 240 (58) 174 (42) 0.19 LIA, n (%) Yes 69 (99) 388 (98) 0.83 43 (98) 407 (98) 0.75 Tourniquet, n (%) Yes 29 (41) 178 (46) 0.49 25 (57) 182 (44) 0.10
Surgical time, min (SD) 89 (23) 87 (20) 0.51 85 (17) 88 (21) 0.33
Side, n (%) Right knee Left knee 42 (60) 28 (40) 193 (50) 195 (50) 0.11 26 (59) 18 (41) 209 (51) 205 (49) 0.27
ASA=American Society of Anesthesiologists, BMI=body mass index, LIA=local infiltration anesthesia, m=mean, NPA=no preoperative anxiety, NPD=no preoperative depression, n=number,
OARSI=Osteoarthritis Research Society International, PA=preoperative anxiety, PD=preoperative depression, SD=standard deviation, %=percent, * = χ2 test for binary data and independent t-test for
Table 5. Change in outcome variables one year after TKR in patient groups without and with preoperative anxiety symptoms.
No preoperative anxiety (n=388) Preoperative anxiety (n=70)
Outcome variable MD (SD) CI p-value * MD (SD) CI p-value MD
between groups p-value** EQ-5D-3L 0.32 (0.32) 0.29–0.35 <0.001 0.33 (0.32) 0.24–0.40 <0.001 0.01 0.89 48 (24) -46 – -51 <0.001 -45 (25) -39 – -51 <0.001 3 0.35 VAS pain
-General health state 11(22) 9–13 <0.001 16 (32) 8–23 <0.001 5 0.17
KOOS-symptoms 30.6 (21) 28–32 <0.001 30 (24) 25–36 <0.001 0.6 0.83 42 (20) 40–44 <0.001 37.2 (26) 32–44 <0.001 4.8 0.08 33 (19) 31–35 <0.001 30 (22) 25–36 <0.001 3 0.18 28 (27) 26–31 <0.001 18 (25) 13–25 <0.001 9 0.006 KOOS-pain KOOS-ADL KOOS-sport KOOS-QoL 43 (25) 41–46 <0.001 38 (24) 33–44 <0.001 5 0.08
ADL=activities of daily living, CI=95% confidence interval, KOOS=Knee Injury and Osteoarthritis Outcome Score, MD= mean difference, QoL=quality of life, SD= standard deviation, VAS=visual analogue scale, *=paired t-test, **=two-way repeated measures ANOVA.
Table 6. Change in outcome variables one year after TKR in patient groups without and with preoperative depres-sion symptoms.
No preoperative depression (n=414) Preoperative depression (n=44)
Outcome variable MD (SD) CI p-value* MD (SD) CI p-value MD
between groups
p-value**
EQ-5D-3L 0.32 (0.31) 0.29–0.35 <0.001 0.33 (0.40) 0.20–0.45 <0.001 0.01 0.86
VAS pain -48 (24) -45 – -50 <0.001 -49 (26) -41 – -57 <0.001 1 0.78
General health state 11.5 (23) 9–14 <0.001 14 (30) 6–24 <0.001 2.5 0.54
KOOS-symptoms 30.5 (22) 28–33 <0.001 31 (24) 24–38 <0.001 0.5 0.89
KOOS-pain 41 (21) 39–43 <0.001 40 (23) 34–48 <0.001 1 0.80
KOOS-ADL 33 (19) 31–35 <0.001 30 (23) 24–38 <0.001 3 0.40
KOOS-sport 27 (26) 21–25 <0.001 18 (25) 11–26 <0.001 9 0.03
KOOS-QoL 43 (25) 41–45 <0.001 37 (23) 31–45 <0.001 6 0.15
ADL=activities of daily living, CI=95% confidence interval, KOOS=Knee Injury and Osteoarthritis Outcome Score, MD=mean difference, QoL=quality of life, SD=standard deviation, VAS=visual analogue scale, *=paired t-test, **=two-way repeated measures ANOVA.
