Author's response to reviews Title: Factors affecting the quality of life after total knee arthroplasties:a prospective study Authors: Ippolyti Papakostidou (pappolly@yahoo.gr) Zoe H Dailiana (dailiana@med.uth.gr) Theodoros Papapolychroniou (thpapa@gmail.com) Lycurgos Liaropoulos (lliaropo@nurs.uoa.gr) Elias Zintzaras (zintza@med.uth.gr) Theophilos S Karachalios (kar@med.uth.gr) Konstantinos N Malizos (malizos@med.uth.gr) Version: 2 Date: 7 May 2012 Author's response to reviews: see over
Author's response to reviewer s comments Title: Factors affecting the quality of life after total knee arthroplasties: a prospective study MS: 9360909896404483 Authors: Ippolyti Papakostidou (pappolly@yahoo.gr) Zoe Dailiana, MD (dailiana@med.uth.gr) Theodoros Papapolychroniou, MD (thpapa@gmail.com) Lycurgos Liaropoulos, PhD (lliaropo@nurs.uoa.gr) Elias Zintzaras, PhD (zintza@med.uth.gr) Theophilos Karachalios,MD (kar@med.uth.gr) Konstantinos N. Malizos, MD(malizos@med.uth.gr) Version: 2 Date: April 30, 2012 Author's response to reviewers comments: see over
2 Prof Serge Perrot Editor BMC Musculoskeletal Disorders April 30, 2012 Dear Professor Perrot, We would like to thank you for reviewing our manuscript MS: 9360909896404483 entitled Factors affecting the quality of life after total knee arthroplasties: a prospective study The reviewers comments and recommendations were taken into account and lead to the revised manuscript that you will find attached. However, the authors of the present study feel that the use of words such as sic from one of the reviewers does not correspond to scientific criticism. In addition, the manuscript was revised from a native-english speaker with scientific expertise. Acknowledgements were included and the sources of funding were added. All changes are marked in the revised manuscript with colour (green for the additions/corrections and yellow for the paragraphs that were re-arranged in the manuscript).according to your instructions we are uploading the cover letter and the revised manuscript through http://www.biomedcentral.com/manuscript/login. We are looking forward to hearing from you soon. Yours sincerely, Zoe H. Dailiana, MD Assistant Professor Department of Orthopaedic Surgery University of Thessalia, Larissa, Greece Assistant professor
3 Author s response to comments of reviewers Reviewer Jan Van der Meulen Comment 1) This is a descriptive study. As a result, it is not immediately obvious what question(s) the authors aim to answer. While reading I thought that the main question was about the differences in POSTOPERATIVE outcomes according to a number of patient characteristics (e.g. age, sex, BMI, social support) while adjusting for preoperative difference. The results could have been better presented. At the moment, there is a table highlighting PREOPERATIVE differences according to patient characteristics and in the text a summary of POSTOPERATIVE differences. I recommend that there is at least one table with POSTOPERATIVE differences according to patient characteristics. Reply: As requested, we provided the 12 months postoperative differences according to patient characteristics in an extra table (Table 4).Concerning the presentation of results, please refer to the response in Comment 2. Comment 2) I also missed in the Results section, a description of the multivariable analyses of the POSTOPERATIVE differences. For example, I would have liked to see what the difference is in outcome(s) between men and women adjusted for the other PREOPERATIVE characteristics. Reply: We apologise if this was unclear; we have now included in the Methods section, the following paragraph: «A general linear model multivariable analysis was used to estimate the effects of all factors of interest (gender age, BMI, level of education, social support and place of residence) on each response variable (QoL questionnaires) at 12 months postoperatively. The corresponding preoperative scores were used as covariates», (page 8, section statistical analysis). Respective conclusions were added in the chapters Results (page 11, 3,4; page 12, 1) and
4 Discussion (page15, 1). In addition table entitled «Table 5. Multivariable analysis of postoperative (12 months) changes in WOMAC domains» was added. Comment 3) There seems to be a discrepancy between the methods as presented in the Methods and Results sections. I didn't see how the presented results were the product of "general linear modelling for repeated measures and post-hoc tests with Bonferroni's corrections". Reply: In the Methods section, we stated: «The effect of each independent variable was analyzed separately for the WOMAC and KSS questionnaires, and the CES-D10 and VAS scores, in time (baseline, 6 weeks, 3, 6, 12 months) using general linear modelling for repeated measures and post-hoc tests with Bonferroni s correction (page 8)».Also, in Table 3, it is stated that Repeated measures GLM with post-hoc Bonferroni s correction. was used. Finally, the results (P-values) in Table 3 and Results section refer to those corrected for multiple comparisons (e.g. in page x we state that: Table 3 also demonstrates that at 6 weeks postoperative follow-up patients did not improve their physical function (P = 0.6) and stiffness scores (P = 0.1).. ). In the GLM, we explored the time effect (5 time points) and we performed pair-wise comparisons between time points. Thus, an adjustment for these multiple comparisons was imposed by applying the Bonferroni s correction. Comment 4) I'm surprised that the definition of "urban" and "rural" is linked to the number of inhabitants in a population rather than the population density (page 5). Reply: According to the Hellenic Statistical Authority (acronym ELSTAT, also known as National Statistical Service of Greece), in Greece a rural areas are considered the municipalities in which the largest settlement has less than 2,000 inhabitants, semi-urban are considered the areas with 2,000-10,000 inhabitants, while in urban areas the population is larger than 10,000. In the 2001 Census, urban and semi-urban areas were incorporated, so we kept this distinction in our study population. (National Statistical Service of Greece, 2008). We have corrected the sentence
5 in the Methods section (page 6, 1), to make it more understandable, as following «In Greece, the municipalities which the largest settlement has less than 2,000 inhabitants are considered rural areas, while semi-urban are considered the areas with 2,000-10,000 inhabitants, and urban the areas with population larger than 10,000». Comment 5) I would have liked to see how many patients did not meet the inclusion criteria. This would help to judge the generalisability of the results. Now we only know that 224 were "eligible" (page 5). Reply: That part was revised as following (page 6, 2): «During the recruitment period 252 patients underwent primary TKA, but 27 did not meet the criteria, 5 refused to participate, and 16 underwent contralateral TKA during the follow-up period and were excluded from the study. The remaining 204 patients, 162 women and 42 men, were included and further evaluated». Comment 6) Some experts would object against the use of face-to-face interviews when collecting the patient-reported outcome measures. Their argument is that some of the instruments used are supposed to be self-administered without any involvement of a researcher or clinician. Reply: According to data published by the National Statistical Service of Greece (2000) 13.6% of the population of 65 years of age had never attended school [1]. Additional in Thessaly the respective percentage is 6.3%, while an additional 8.4% has not finished the elementary school [2].When this data was take into account we planed face to face interviews to facilitate mainly the older and illiterate patients. The involvement of the same interviewer in every case minimized the bias. [1] National Statistical Service. Level of Education 2000 Census. 2000. Retrieved from http://www.statistics. gr/portal/ page/portal/ver-1/esye/bucket/a0101/ Other/A0101_ SJO01_TB_ QQ_ 04_2000_ 01A_F _BI_0.pdf. [2]N Fakiolas: Reasons and Policies Towards Illiteracy The Greek review of Social Research 2006, 120: 169-194.
6 Comment 7) Level of education is dichotomised into "primary" or "less/ otherwise". I thought that this was unclear (page 7). Reply: We agree with the reviewer that used terms are unclear and we decided to replace them throughout the text and in the tables with the terms «elementary/ less» and «high» (page 8, 3 and Tables 1,2,4,5). According to the data presenting in comment 6 we decided to separate (split) the population into two groups and the border-line was the elementary school. Comment 8) Bonferroni s correction for multiple comparisons is mentioned. How was this applied and how is the correction represented in the description of the results? Reply: Please refer to the response in Comment 3. Comment 9) There are a lot of detailed figures included in the text of the Results section that repeat results already presented in the tables. I would have preferred to see the headline results in the text helping the reader to digest the tables rather than a repetition of actual figures. Reply: According to the reviewers suggestion we omitted table 2 and also omitted from the text, data presented on table 1. Comment 10) In the Discussion, the pattern of improvement of outcomes over time is highlighted (greatest improvement in first 3 months with smaller but ongoing improvement thereafter (page 11). This demonstrates again that we need to know what the question(s) are that the authors aim to answer: differences in outcomes according to preoperative patient characteristics or pattern of recovery of outcomes according to time after operation. If both these questions are relevant I would separate them out and deal with them in two sections. Reply: The main question is the influence of patients characteristics on outcome. The pattern of recovery is a separate finding, very useful to the clinician, from our point of view. However, we
7 would not like to add a second question to our main research question. Thus the 2 nd of the discussion, highlighting the improvement of outcomes, was moved after the discussion of our primary research question and is now 7 (page 14). Comment 11) The authors also mention in the Discussion that BMI does not have a significant effect on outcomes. I would have liked a discussion about the power to detect his effect. The study population is small and dividing patients in two groups (BMI smaller or larger than 30 kg/ m^2) may not be the best way of investigating the impact of body mass. Reply: In the Discussion, (page 13, 1) we included the following sentence «However, the power to detect a significant BMI effect on pain and functional limitations at 12 months postoperative follow-up was limited (power=12% and 23%, respectively), and therefore, the results should be interpreted with caution. It must also be noted that patients with morbid obesity (BMI > 40kg/m 2 ) were not offered the option of TKA from the surgeons of the present series, during the study period». The study population might be considered relatively small; however, the distribution of the population using as threshold the value BMI=30 (which is typical threshold value) allowed us to investigate the BMI effect. The 2 subsets according to the BMI variable (<30 and 30-40) comprised of quite symmetrical samples, therefore comparisons were easy. Comment 12) In the Discussion, a number of issues were brought up and in some cases I thought that the authors were at risk of over interpreting their findings. An example in this context is the discussion of the relationship between depression and pain after the operation (page 13). Reply: The pre-operative relationship of depression and pain is supported by the findings of other studies. To avoid over interpretation of postoperative findings we rephrased the respective paragraph (page14, 2).
8 Reviewer Montserrat Núñez Comment 1) The written English of the article is poor and does not meet the standards of the journal. Apart from various obvious mistakes Would infection (Table 1), radiographic sings (page 11), the phrasing is clumsy, with the wrong verbal tense used, often making interpretation difficult. The article should be revised by a native English speaking medical translator in order to make it more comprehensible, without ambiguities, and with a better style which could probably reduce the article length. Reply: The article was revised by a native-english speaker with scientific expertise. Comment 2) The authors first state that patients were included if they were speaking Greek but later state For patients unable to read Greek, the questions were read out by the interviewer. This is confusing are we talking about immigrants or illiterate subjects? Please clarify. Were any patients excluded for these reasons? Reply: The high rate of illiteracy in the elderly forced us to design the data collection accordingly. According to data published by the National Statistical Service of Greece (2000) 13.6% of the population of 65 years of age had never attended school [1]. In Thessaly the respective percentage is 6.3% while 8.4% had not finished the elementary school [2]. Patients that could not speak Greek were excluded from our study. [1] National Statistical Service. Level of Education Census. 2000. Retrieved from http://www.statistics. gr/portal/ page/portal/ver-1/esye/bucket/a0101/ Other/A0101_ SJO01_TB_ QQ_ 04_2000_ 01A_F _BI_0.pdf. [2]N Fakiolas: Reasons and Policies Towards Illiteracy The Greek review of Social Research 2006,120: 169-194.
9 Comment 3) The authors state that patients with osteoarthritis were studied, but later state that 4% of patients had posttraumatic arthritis. What is the rationale for this? Reply: The reviewer is right. We included patents with arthritis (OA, traumatic) and we excluded all cases of inflammatory arthritis due to the multi-joint involvement that could influence the outcome. The phrase «they suffered from severe knee arthritis (OA or traumatic)» was added in page 6 1 of the manuscript. Comment 4) The authors do not state exactly how many patients were included in the final analysis. I see that two patients died in the early postoperative period Were these patients included in the baseline or later analyses? Reply: A total of 204 patients were included in this study. Two patients died at the early postoperative period and were excluded from the postoperative evaluations. Twelve months postoperatively the response rate was 90.2%. Comment 5) The authors state that the global mean age was 69.17 years. However, later they state that 86% of patients were discharged directly at (sic) home, and had a mean age of 65.6 years, and that the remaining 14% were transferred to rehabilitation centers and had a mean age of 68.9 years. These figures seem contradictory. Please clarify. Reply: Yes. You are right.the correct figure was 69.6 years, instead 65.6 years. We rephrased the respective paragraph of the manuscript (page 9 5). Comment 6) The authors state that patients were classified according to urban (>10,000) or rural (< 2,000) status. They do not state anything about the patients coming from areas with 2-10,000 residents. Semi-urban or semi-rural? Please clarify. Also, it would be of interest to know how many patients came from each hospital and whether there were any differences between them and
10 what the catchment areas are for each hospital. Likewise, I believe that Larissa is a fairly large city and therefore it would be of interest to know how the rural/urban status was determined. By municipality? Reply: According to the Hellenic Statistical Authority (acronym ELSTAT, also known as National Statistical Service of Greece), in Greece rural areas are considered the municipalities in which the largest settlement has less than 2,000 inhabitants, semi-urban are considered the areas with 2,000-10,000 inhabitants, while in urban areas the population is larger than 10,000.. In the 2001 Census, urban and semi-urban areas were incorporated, so we kept this distinction in our study population. (National Statistical Service of Greece, 2008). We have also corrected the sentence in the Methods section (page 6, 1) to make it more understandable, as following «In Greece, the municipalities in which the largest settlement has less than 2,000 inhabitants are considered rural areas, while semi-urban are considered the areas with 2,000-10,000 inhabitants, and urban the areas with population larger than 10.000». In total, 26 patients came from the hospital in Athens and were classified as urban population. The remaining patients (176), coming from the region of Larissa, were classified as urban/semi urban or rural inhabitants. Comment 7) The authors state that 44.5% of patients were depressed preoperatively according to the CES-D10. Later they state that depression was not detected at (sic) any of the examined groups during the post-operative follow-up intervals. Later, in the Discussion, they state that 7.35% of patients remained in a depressed mood at 12 months. Please clarify these apparent contradictions. Reply: We agree with the reviewer. The sentence was rephrased as following: «High scores, indicative depression were not found among the different groups of the study, during the
11 postoperative follow-up» (page 12, 1). The small percentage of patients (7.35%) with depressed mood 12 months postoperatively did not affect the final outcome (values 10 in CES-D10) and comparisons between the study groups (Discussion, page 14, 2). Comment 8) The authors state in the Discussion that patients with BMI>40g/m2 were not included in the study. However this is not stated as an exclusion criterion. What was the rational for this, and how many patients were excluded for this reason? Reply: Patients morbidly obese were not excluded from our study. Simply, these patients (BMI > 40kg/m 2 ) were not offered the option of TKA from the surgeons of the present series, during the study period. We rewrote the sentence in the manuscript to clarify this (page13, 1). Comment 9) The authors state that one investigator administered the questionnaires and evaluated parameters including range of motion. However, this investigator does not appear to have any medical or other qualifications according to the Title page. Please clarify what training this investigator has or received. Reply: The investigator is a registered nurse and PhD candidate. Comment 10) There are various ambiguous statements. For example, in the description of the KSS in the Methods section, the authors state In addition, independent assessments of depression and pain were conducted at each evaluation to determine what psychological benefits patients are likely to experience after the surgical intervention. What does this mean exactly? Likewise, the authors state that 86% of patients were discharge directly at (sic) home, the majority of whom (93%) received physical therapy services at home from (the ) physical therapist of their choice (12 sessions of physiotherapy in 6 weeks, starting the day after discharge. Apart from the
12 fact that the 86% and 93% figures are reported in Table 2 without distinction, what does this sentence mean? Were patients able to choose their own physio? Who provided physiotherapy? Equally, the authors should describe what they mean exactly by rehabilitation centers for readers accustomed to other types of healthcare models. Reply: A.We rephrased the manuscript (page8, 1) as following «potential improvement of mood». B. Table 2 was removed and the manuscript was revised as following «The majority of the patients (176) went home after discharge and followed a rehabilitation program supported by a physiotherapist with 12 sessions of physical therapy in 6 weeks period, starting the day after discharge. The remaining 28 patients were transferred to rehabilitation centers with a mean length of stay 19.3 ± (3.3) days and received a similar program of physiotherapy. Most patients (176) went home after discharge and followed a rehabilitation program supported by a physical therapist (12 sessions of physical therapy in 6 weeks, starting the day after discharge). The remaining 28 patients were transferred to rehabilitation centers and received physical therapy services; mean LOS 19.3 ± (3.3) days» (page10, 2). C. Rehabilitation centers in Greece have similar infrastructure and organization as in other European countries and are subject to the supervision and control of Ministry of Health (www.yyka.gov.gr). Comment 11) In fact, Table 2 provides little information and could be eliminated, with the results included in the text. Reply: This table was eliminated Comment 11) The references seem a littel old and could be updated quite easily.
13 Reply: According to reviewers suggestion we added the following references: No 14,22,24,26,28,31,36,43,44,47,48.We also removed the references No 11,29, 33 (first numbering) Comment 13) The authors should state the minimum clinical difference used to estimate the power of the study and therefore the sample size needed. Reply: Our study was an observational prospective explanatory study with a predefined follow up of 12 months. Therefore, a formal calculation of sample size was not performed. However, by evaluating 184 subjects at the end of follow up, the power of detecting the observed difference between preoperative and postoperative (at 12 months) score in WOMAC pain (which is δ=10.2) and considering the observed variability (which is σ=4.4), is estimated to be 99%. Comment 14) The authors should analyse the effect size of the improvements in QoL. Reply: In Methods, statistical analysis section (page 8, 3), the following sentence was added (page: «Effect sizes (ES) were calculated for the different outcome measures using the formula ES=mean change/sd of preoperative scores.the effect size is a standardised measure that provides information regarding the magnitude of change before and after TKA. An effect size of 0.8 or greater is considered large (Nunez et al. 2009)».In Table 3, the ES was added. Comment 15) The authors state a hypothesis in the Discussion. If indeed this is the basic hypothesis underlying the study, it should be stated in the Discussion. Reply: In the first sentence of the discussion the authors state the same research question of the last paragraph of the Background section Comment 16) The desciption of the surgical characteristics does not seem to correlate with the variables described in the Methods. Some of this information may be superfluous.
14 Reply: We agree and eliminated these characteristics. Comment 17) WOMAC data should be normalized to a 0-100 scale if possible. If not, the tables should contain the values of the dimensions as a foot note. Reply: This was done. Comment 18) Abbreviations in the tables should be described. Reply: This was done. Comment19) Figure 2 is not clear. Reply: This Figure 2 was removed. Comment 20) In table 5 the authors state 1.5 months. In the rest of the article, 6 weeks. Please unify. Reply: We replaced the 1.5 months with 6 weeks in table 5.