ࡱ> ikhmq Jbjbjt+t+ KAA|]BBBVVVV8Vjj(/%Ty,ؗڗڗڗڗڗڗ$ɘFB "/u eju u u 8BؗVVؗu u A *BؗF$$VVg 4. Discussion and conclusion 4.1 Introduction In this chapter a discussion and conclusion was conducted around the results found for this review. This will help to understand the implications of the evidence in relationship to practical decisions about healthcare or future research. In the first section we will address the important methodological limitations of the included studies. In the next section we will give an interpretation of the results obtained. In the third section the limitations and the implications for research and practice will be discussed, and in the final section we will give a conclusion based on our findings. 4.2 Methodological quality 4.2.1 Internal validity For the purpose of methodological quality assessment of included trials, a decision was made to use the PEDro scale. The Pedro scale is a quality assessment tool based on the Delphi list. The decision to use PEDro was primarily due to the easy-administered nature of PEDro scale, and because an informative tutorial on the administration of the scale was available on the PEDro website. The score obtained by the included studies on the PEDro scale ranged from two to six, out of a maximum score of ten. There were certain methodological characteristics that could be derived from the quality of the included studies. Figure 4.1 shows each of the 11 items on the PEDro score, and the number of articles that managed to obtain a point for each of these items. There are certain methodological characteristics we can extract from the included trials by studying figure 4.1: -No studies have been able to blind subjects, therapists or assessors -Only one study reported intention to treat analysis -Only four studies reported concealment of allocation. -Only five studies managed to obtain adequate follow up -Nine studies had groups with similar baseline values -10 studies were randomised-controlled trials. -10 studies provided both point measures and measures of variability -11 studies specified eligibility criteria for inclusion of subjects -All studies reported between-group comparison results. Of the ten studies that were randomised controlled trials, only four of them reports concealment of allocation. This is a major source of bias in the randomisation procedure (39), and reflects the poor methodological quality of many of the studies included in this review. The fact that only one author (8) reports intention to treat analysis is most likely due to poor reporting. There were few participants in these studies, and any bias in allocation is not very likely. However, this was not reported in all but only one study, so this indicate a weakness in the reporting of the authors Figure 4.1 Showing the number of studies ( total n=12) fulfilling the items (1-11) on the PEDro score: Items: 1.Eligibility criteria were specified; 2.Randomization; 3.Concealed allocation; 4.Similar baseline values between groups; 5.Blinding of subjects; 6.Blinding of therapists; 7.Blinding of assessors; 8.Adequate follow up; 9.Intention to treat; 10.Results between-group statistical comparisons are reported for at least on key outcome; 11.Point measures and measures of variability for at least one key outcome. During the personal communication phase of this project, contact was obtained with Dr. Arianne Verhagen. She is one of the developers of the Delphi List, and recommended to use this scale instead of the PEDro scale. She justified her preference for the Delphi List for two reasons: 1. The Delphi List is one of the few validated quality assessment tools and therefore recommended by the Dutch Cochrane Centre, and 2. The PEDro added two statistical items to the Delphi, so the relative weight of the statistical items in the overall quality score increases at the expense of the weight of the internal validity items. Contact was also obtained with Dr. Rob Herbert of the Centre for Evidence-Based Physiotherapy. The Pedro scale was developed at this centre, and Dr. Herbert argued favourably for the use of the PEDro scale. No assessment tools for trial quality have been found to be a golden standard. Different measurement tools focus on different aspect of trial quality. To make an interesting point regarding methodological quality assessment, a decision was made to assess the included studies with the Delphi List, and compare the results with the results from the PEDro scale. Table 4.1 shows the scores obtained by the 12 included studies on the Pedro scale and the Delphi List. Table 4.1: Shows the score obtained by the 12 included studies on the Pedro scale and the Delphi list. Scores are given in both point scores and percentage of maximum. StudyPEDro scoreDelphi scoreScorePercent scoreScorePercent scoreSuomi 20005/1050%4/944,4%Green 19935/1050%5/955,6%Ahern 19955/1050%5/955,6%Sylvester 19893/1030%3/933.3%Hall 19966/1060%5/955,6%Nguyen 19976/1060%5/955,6%Helliwell 19963/1030%3/933,3%Baldwin 19782/1020%2/922,2%Minor 19894/1040%4/944,4%Stenstrom 19913/1030%1/911,1%Hansen 19935/1050%4/944,4%Smith 19984/1040%4/944,4%TOTAL46/12038,3%45/10841,7% Apart from the study of Stenstrom (1991) that scored 30 % on the Pedro Scale but only 11,1 % on the Delphi List, there were no major differences in the scores obtained by the included studies on the two scales. Stenstrom was the only author who did not report eligibility criteria in his trial report, and this seem to be the reason why he scored so much lower on the Delphi List. The means scores of the included trials were 38,3% on the Pedro scale, and 41,7 % on the Delphi List. This difference of 3,4 % indicates that the scales show similar results. A more interesting note on the mean scores is that they fall way below 50 % on both scales, which indicates that there is a greater chance that the studies are biased to some extent, than that they are not biased. This fact implies that caution should be taken when interpreting the included studies, as the results from them are likely to be invalidated by methodological flaws. Only six of the included studies scored 50 % or higher on the PEDro scale. This needs some explanations: First, the amount of research conducted on aquatic therapy is scarce, and sadly, the quality of many studies is simply just poor (26). Second, it lies in the nature of aquatic therapy that some criteria on the scale may be impossible to fulfil. None of the studies scored points for blinding of subjects, therapists or assessors. Blinding of subjects is virtually impossible due to the fact that subjects will know whether they have been immersed in water and performed exercises, or not. Therapist will also know whether they have been instructing aquatic therapy exercises to patients, or not. The criterion blinding of assessor was considered to be possible. However, according to the tutorial on the administration of the PEDro scale, it is explicitly stated that if key outcomes are measured by self-reported measurement tools, assessor is only considered blinded if the subject is blinded (40). The impossibility of blinded subjects in trials on aquatic therapy has already been mentioned. It must also be mentioned that two of the outcomes being investigated in this review, pain and HRQL, are usually only measured by self-report. Hence, blinding of assessor, therapist and subject may be impossible for these two outcomes in relation with aquatic therapy. Blinding of the assessor was considered possible for measurements of ADL-functioning, whenever objective measurement tools were used (e.g. timed walking tests, ability to rise from chair-tests and stairclimbing-tests) However, in the studies included in this review , function was primarily assessed by means of a self reported questionnaire, The Stanford Health Assessment Questionnaire (HAQ). Therefore, no studies managed to obtain a point for blinding of assessors. Another note on blinding of assessor is that in case of objective measurements being taken, there is a great probability that participants will reveal to the assessor what group they are allocated to. Most likely some verbal interaction will take place between the subject and assessor, so at least there is a great possibility of this happening. Only one of the included studies reported that subjects were specifically told not to reveal the assessor which group they had been allocated to (11) Due to these limitations, the highest PEDro score a study could obtain was considered to be eight out of 10, as blinding of subjects and therapists was considered impossible. These limitations will also be present in other trials on active physiotherapy treatment. A subject performing exercises will never be considered blinded. The same goes for blinding of therapists, as they too will know whether or not they are instructing exercises or giving placebo intervention. These limitations will be present in all kinds of active rehabilitation interventions. Blinding therapist and subject is fairly easy in medical drug interventions, as the active drug and the placebo-drug can be designed to look, taste, smell and feel identical. Research within the field of rehabilitation medicine will therefore always have limitations that medical, and many other sciences, will have the possibility to eliminate. A possibility could be to develop a scale for rehabilitation trials, where these limitations are taken into consideration (see 4.5.2). In this review, where blinding of therapists and subjects was considered impossible, these criteria could be removed from the scale to give a more realistic view on the methodological quality of the trials. The quality scores obtained if these two criteria were removed can be seen in table 4.2. Table 4.2: Methodological quality scores as it would have been if the criteria blinding of therapist and subject is removed from the PEDro scale.( max. score is eight points). Author / yearPoint score% score of max. scoreSuomi 20005/862,5 %Green 19935/862,5 %Ahern 19955/862,5 %Sylvester 19894/850 %Hall 19966/875 %Nguyen 19976/875 %Helliwell 19963/837,5 %Baldwin 19782/825 %Minor 19894/850 %Stenstrom 19913/837,5 %Hansen 19935/862,5 %Smith 19984/850 % After trial quality assessment had been conducted by use of the PEDro scale, some of these trials were identified on the PEDro database. On this database, most trials have been assessed for quality by this scale. The official PEDro ratings were considered golden standards, so the official ratings were compared with the ratings given by the authors of this review. On some ratings, there were discrepancies between this project groups rating and the official PEDro ratings. These ratings were re-assessed. A decision was made to dispute some of the official ratings (see appendix V for personal communication with Dr. Rob Herbert regarding dispute of PEDro ratings). Dr. Herbert, director of the Centre for Evidence based Physiotherapy, agreed with this project groups opinions on some ratings. These official PEDro ratings were subsequently changed. 4.2.2 External validity Regarding the external validity, or the generalisability of the included studies, a few points need attention. There were 785 participants in the 12 studies. Of these, only six percent were diagnosed with ankylosing spondylitis. Only one study that was considered eligible for this review dealt with ankylosing spondylitis (6). This study did not consider aquatic therapy on its own, but rather as a component of a total physiotherapy regime involving exercises and massage. 46 percent of the total number of participants were diagnosed with OA, and 48 percent with RA. These uneven percentages are in accordance with global prevalence statistics, where RA and OA are far more frequent than AS. (see 1.3) Of the studies giving sex ratios, there were in total 67 percent women. This sex ratio is very similar to the RA and OA population in general, where women are affected more than men (see section 1.4.1). Helliwell (2000) and Baldwin (1978) dealt with participants with a mean age of 44 and 11 years, respectively. The remaining ten studies had a mean age of their participants ranging from 52 years to 67 years. Except from Baldwin (1978), no other studies have been conducted on younger populations. Considering that the average disease duration of all participants being nearly 12 years. A description of the main population the results from this review are applicable to, can be done to get a clearer picture of the generalisability of the results. In broad lines, the research that has been conducted on aquatic therapy, has investigated its effect primarily on women between the ages of 52 to 67 years, with a diagnosis of RA or OA for 12 years. Interpretation of results 4.3.1 Discussion on pain Pain was one of the outcomes that was of interest for this review. Comparing the results on pain directly was difficult due to the lack of standardised measurement tools and differences in design of interventions, patient population and sample size (see 4.2). However there are some occasions interesting relationships between intervention and outcomes. As stated in section 3.4, there were 7 studies reporting reduction of pain for the aquatic group at post treatment, and two studies reporting no effect at all. None of the included studies reported increase of pain at post treatment. The two studies reporting no effect (Hansen and Stenstrom) used long term treatment (2 and 4 years respectively), while all three studies using short term treatment (Nguyen, Hall and Helliwell) reported reduction of pain. All short term treatments lasted for three weeks. These result could be due to many factors. One of the authors from the long term treatments (Hansen) explains that the positive results in short term treatment studies, can be explained by the possibly greater enthusiasm of the patients and health professionals which again can influence the results in a positive way. Hansen also states that there was a small sample size in each group in the study, which increases the possibility of chance effects. The other long term study (Stenstrom) had a low score on the PEDro rating (score 3) which indicates a weakness of evidence and hence, makes it too difficult to draw any conclusions from this study. It is interesting to notice that studies performing high frequency treatments achieving better improvement on pain than the low frequency treatments. Only one of the low frequency treatments reported improvements pre- to post treatment (2). However this study investigated improvements for children with juvenile rheumatoid arthritis and can not be compared to the other studies. The lack of positive results in the low frequency treatments are often associated by the belief that too low intensity has no benefits, as also suggested by Hansen. Another interesting point to notice is that most studies report improvements on pain for both aquatic and land based exercise groups. Only the studies of Baldwin and Helliwell report improvements for the aquatic therapy group alone. Even though the land based exercise groups also showed improvements, the aquatic group reveals greater improvements. This correlates with the beliefs that the component of warmth and water has valued benefits on pain. When looking at the study of Hall, the immersion group also showed reduction in pain, but the aquatic group and the land based exercise group showed greater reduction of pain, which indicates that all components of aquatic therapy can reduce pain. Of the five studies giving follow up data, three studies (Nguyen, Ahern, Minor) reported maintenance of improvement for the aquatic groups (RA and OA). As rheumatism is a chronic disease, the duration of the treatment effect is of great importance, therefore follow up testing should be done in all studies. 4.3.2 Discussion on ADL functioning ADL function is one of the most important outcomes to evaluate nowadays (1.4), the medical world agrees that it needs to be included in future aquatic therapy exercise program evaluation. It seems very important to people with rheumatic conditions (as to any other kind of patients) to keep a stable ADL function. In this review, the report of the results of aquatic therapy effects on ADL functioning shows that the 7 authors share different opinions on the changes The authors of this review have first noticed that none of the researchers included focused on functional exercises as such. It is not the issue to discuss here if exercises in water can actually be functional in the way they are done (exercises where patients train to walk upstairs or put on shoes in water are rarely seen!). But it has been seen that the exercises in water chosen had often as objective the improvement of some impairment: strength, pain, ROM, aerobic capacity with an eventual beneficial consequence on ADL functioning and HRQL. If it can be considered that Hall and Suomi's studies are more reliable because of a slightly higher methodological quality (6 and 5 respectively), so their results should maybe represent more weight. Both of them conclude of some improvement in functional measures following aquatic therapy. Both of them include people with RA on a "twice" or "three times a week" treatment, this for four to six weeks. Those intervention parameters are not to be taken as the only ones to produce effects. Indeed, Suomi evaluate status through a measurement of lateral sway, which is a component of gait quality. But it is not a way to conclude that ADL function is better if lateral sway is improving. This illustrates that maybe ADL function measure has to gain with some standardization in the tools being used. Hansen and Stenstrom are the two "long term treatment" authors (two and four years). Considering the cost of aquatic therapy for the society, long term treatment studies are important to perform. Are aquatic therapy effects worth the cost? It is difficult to answer to that question. Hansen (five on PEDro scale, three times a week treatment) mentions contradicting facts: patients reports subjective improvements, which are not confirmed scientifically in Hansens study. Both Hansen and Stenstrom study on people with class II RA. This has maybe, as well its importance and it is questionable that people more severely handicaped (III or IV on Steinbrocker) could follow such long-term treatment! In favour of long term treatment for RA patients, one can maybe think that it is actually "unfair" to conclude that "no significant effects" is not good enough for such category of patients. It needs then to be reminded that in a two or four years long treatment period, the situation of a person with RA is probably degrading. Then the effects of aquatic therapy are indeed greater than tests actually show if the disease is active while treatment is going on. The maintenance of a functional status can maybe be seen as a positive effect. One other important point to consider is the importance of aquatic therapy effects compared to other physiotherapeutic means. In view of the four studies opposing aquatic therapy with exercise therapy, the observation of the results maybe suggests that aquatic therapy treatment is not more effective than exercise therapy on land for ADL function (see table 3). This is a rather puzzling conclusion if it is true! Smith for example, with 24 RA patient concluded to "no significant difference between aquatic therapy group and control group". Since the amelioration is similar in both treatment and control group (ROM program), it can hardly be said that aquatic therapy effects prevail on exercise therapy effects. The same statement can be said when comparing aquatic therapy effects with immersion "alone" effects on ADL functioning. Hall suggests that immersion alone, during a four week long treatment is sufficient to improve ADL function. Logically, the positive effects of exercise after an active aquatic therapy treatment (see 1.2.3: effects of resistance and turbulence plus heat and hydrostatic pressure), could add to those of an "immersion alone" treatment (use mostly of heat and hydrostatic pressure). But Hall is the only author doing this comparison, so conclusion is to be taken carefully. It is important to note that the study done by Helliwell is the only one done on people with AS, and that it does not include an ADL functioning evaluation. No results can therefore be showed on ADL functioning improvement for AS patients in this review. The summary proposed after the analysis of the studies on aquatic therapy effects on ADL functioning is richer in question marks than conclusions. It seems that aquatic therapy is effective on ADL functioning for RA patients following a short program (up to six weeks) with a quite high intensity (three times a week). But nothing is sure about the better effects of aquatic therapy compared to some forms of exercise therapy or immersion alone. Nothing is known either for people with AS or JRA and about the possible relation between effects and parameters such as age, sex or duration of disease. Long term treatments are maybe beneficial if the maintenance of ADL functioning in a two or four years program can be seen as a positive effect. 4.3.3. Discussion on HRQL. HRQL is an outcome measure that consists of many variables representing the general satisfaction of a person. Depression, anxiety, mood, tension, social activity and pain are among factors that are measured in a widely used HRQL measurement tool, the AIMS (see 1.4.1). Many studies have not included HRQL as an outcome measure, and therefore are prone to miss out important aspects that can be related to disease symptoms. In the study of Hansen, there was a discrepancy between the lack of treatment effect on the variables measured (pain, swelling, x-ray, ESR and Hb) and the fact that most patients generally felt improvement. Hansen suggests that an outcome measure concerning HRQL may have been better suited to detect the general satisfaction of the patient. Of the twelve studies included in this review, only five measured HRQL. Of these, four studies showed improvement on HRQL. The one study showing no improvements was the study of Hansen (1993), which have been discussed earlier in this chapter. As explained in 4.3.1 Hansen justifies many reasons for the lack of results. If not considering the study of Hansen, aquatic therapy seems to have a positive effect on HRQL. Improvement on HRQL has often been noticed where people experience social benefits of group therapy interaction, and in this particular case, the pleasure of being in warm water. Hall concludes saying that a combination of water and exercise showed that aquatic therapy gave superior benefits in terms of physical and psychological functioning compared to other interventions. As the psychological factors are important for quality of life, this may explain the valued effects of aquatic group therapy on HRQL. As opposed to the follow up testing on pain, all studies reported maintenance or improvement on HRQL for the aquatic group. This is of great value for patients having a chronic disease as rheumatism. The summary proposed after the analysis of the studies on aquatic therapy effects on HRQL can be simply done: a majority of authors (four out of five) evaluating HRQL agree in saying that the psychological beneficial effects of group aquatic therapy is positively influencing HRQL. 4.4 Conclusion Research on aquatic therapy for patients with rheumatoid arthritis, osteoarthritis and ankylosing spondylitis is at this point in time inconclusive regarding the effects of this intervention on pain, ADL-functioning and Health Related Quality of Life (HRQL). This is especially the case for ankylosing spondylitis, where only one study that incorporated aquatic therapy was considered eligible for this review. Few studies have been able to demonstrate positive effects on objectively measured outcomes. However, many authors state that one can not ignore the positive subjective responses from patients participating in such programs. There is a clear tendency that subjectively measured outcomes, like pain, self-reported function, life satisfaction, depression, mood, tension and HRQL improved more than objectively measured outcomes. Short-term improvements in pain (1,2,4,6), and long term improvements in HRQL (4,7,8) were reported by several authors. An objective gain in functional status was not likely. This is most likely a consequence of the fact that none of the aquatic therapy programs implemented in the included studies, related exercises to the patients functional limitations. In other words, aquatic therapy may have more of a psychosocial effect than a physical one. Studies included in this review differed in study design, outcome measures and methodological quality. No statistical pooling could be performed on their results due to this heterogeneity. The lack of objective and statistical evidence is challenged by the patients satisfaction with this intervention. Several authors found some justification for the continued use of aquatic therapy. An important point is that no authors reported harmful effects of aquatic therapy, so in the worst case investing in this treatment form is merely a waste of resources. 4.5 Implications 4.5.1 Implications for future practice. Two questions can be posed when discussing the implications this review has for future clinical practice. The first can be posed to referring physicians, and deals with what kind of people will benefit from aquatic therapy. The other question can be directed towards providers of aquatic therapy, i.e. physical therapists. This question deals with the nature of the therapy. What intensity, treatment frequency, type of activities and duration of therapy is optimal for people with RA, OA and AS. With only one study to back up effects of aquatic therapy for ankylosing spondylitis, no implications for future practice can be made for this population. For the RA and OA population, some implications can be made. The primary effects of aquatic therapy have been found to influence psychological variables more than physical variables (1,4,5). Pain and HRQL are more likely to be influenced than ADL-functioning. This suggests that people with RA or OA that experience severe pain or who report dissatisfaction with their life quality, may benefit from this intervention. It can not be expected that people with RA or OA who experience functional limitations in their activities of daily life will benefit from aquatic therapy. However, no studies have reported harmful effects of aquatic therapy. Therefore, this is an active treatment form that can be implemented in phases of increased pain when other weight bearing activities are impossible. For the providers of aquatic therapy a few implications for future practice can be mentioned. Aquatic therapy is not harmful. Even studies on intensive aquatic therapy programs with three or more sessions per week have not shown any negative effects. General exercises aimed at increasing strength, mobility and aerobic capacity may have positive short term effects on pain and long term effects on HRQL. An improvement in ADL-functioning is less likely with aquatic therapy programs implementing general exercises. 4.5.2 Implications for further research. The methodological quality of the included studies was poor. This can be explained by mentioning that aquatic therapy faces similar problems as other rehabilitation interventions when it comes to blinding of therapists and participants (26). A question to answer for future researchers within the field of rehabilitation, is if any benefits can be obtained by producing a quality assessment tool that takes these things into consideration. Researchers investigating effects of a treatment for chronic rheumatic disorders should focus on the evaluation of outcomes that are of importance for the patients . Pain, ADL function and HRQL have been identified as the three most important outcomes for patients (28). Investigating effects of therapy for rheumatic patients means to pay attention to long term effects. Patients with chronic rheumatic diseases have usually a long disease history, not only behind them but as well ahead of them. Post-treatment assessments should also include long term follow up evaluations of effects. No studies have yet investigated the effects of aquatic therapy on people that have been diagnosed with a rheumatic disorder for only a short period. Early intervention aquatic therapy programs may prove to be more effective than programs that are implemented many years after a diagnosis have been established. Future studies should address this problem. No research has been conducted on individual functional aquatic therapy. The lack of objective positive findings of aquatic therapy on ADL-functioning can be due to the general nature of the exercises performed. Individually prescribed exercise programs in relation to the patients specific functional limitations can perhaps be a better suited treatment method for people with limitations in their ADL-functioning. This needs to be investigated in future studies. In general, studies that have been conducted on the effects of aquatic therapy have been of poor methodological quality. To provide more conclusive evidence regarding effects of aquatic therapy, researchers should attempt to conduct high quality studies. This involves performing controlled trials with concealed randomisation. In addition the following elements should be described in future studies: baseline characteristics, drop-outs, blinding procedures, outcome measures, assessment tools, intervention characteristics, co-intervention characteristics. If all these elements are described better in future studies, a clearer picture of potential bias can be obtained. It could be suggested that some research board would not only evaluate the product of research but also synthesise it in order to produce guidelines (this would maybe be easier with some kind of standardisation in todays research) . Such initiative would prevent that research is pulled in different directions and make the establishment of guidelines so difficult. Limitations of this review The process of conducting systematic review proved to be quite difficult for the members of the group considering the limited experience we had in the field of research. Quite some of the time available for this project was spent on broadening the knowledge on how to conduct a systematic review properly. Had we been familiar with the Cochrane reviewers handbook beforehand, time could have been spent more efficiently on the actual process of conducting a systematic review. Especially challenging was presenting and interpreting the results from the trials we found. None of us had any previous experience in statistical analyses, so any real interpretation of the results was beyond our base of knowledge. Further on we realised that our research question was too broad considering the limited time available for this project. This made it difficult to go in-depth into the interpretation of the results from the review. The studies that were found differed widely in methodological quality, and participant and intervention characteristics. The different studies also focused on different outcome measures. This made the interpretation of the results difficult for the reviewers. PAGE 1 PAGE 36  EMBED Excel.Sheet.8  &):4 H J  24CD[\WYghn2<P\pܾܾB*mH  56B*mH 6B*6CJ 56CJmH  6CJmH  B*CJmH jCJUmH  5CJmH CJmH CJhmH nH5CJhmH nH 5CJmH  5CJ,mH A&'():; 4 5 { ! W  H H&d &'():; 4 5 { ! W  H I J ghnz#'+12<AEIOP\a   ^H I J ghnz$$$z_|\$X$$F t     r      $@$$F t     F    $$ #||$X$$F t     r      $$#'+12<AEIOP\aeiopx$X$$F t     r      $$aeiop !&*.45;BHOUVWX6#7#8#9#:#+%,%((g*h*i*++*+6+M+N+Y+]+d+e+p+t+{+|+++++++++++++++++++   _|$X$$F t     r      $$!5;VXi*s*+M+N+d+e+p+{+|+++++++++++++++++,,,,,,,7,8,D,O,P,[,d,e,////e666BBBBWWWi`k`z`{`eʽĽ 5CJmH 55CJmH  5CJmH CJmH  B*CJmH  6B*CJ56B*CJmH  5B*CJ 56CJ6CJ 56B*mH B*mH G $X$$F t     r      $$ !&*.45;BHOU|$$$X$$F t     r       UVWX6#7#8#9#:#+%,%((g*h*i*+X$$F t     r                ++*+6+M+N+Y+]+d+e+p+t+{+\r\o$@$$F t     F3      $$$@$$F t     F3  $ {+|+++++++++++++++++++++\dPXl\$$$@$$F t     F3      ++++,,,,,,,,,0,7,8,D,H,O,P,[,_,d,e,f,g,h,///////L0M022244c6d6e6666678<>cABBBBBCDDFFIIKKLLNZPQQSSTTWWWWWXYZZ^O_i`j`k`z`{`eg 0   ]++,,,,,,,,,0,7,8,D,H,O,Tl`@$$F t     F3      $$$O,P,[,_,d,e,f,g,h,/////Tusssssssq@$$F t     F3       $$$B$$F t     F3       ///L0M022244c6d6e6666678<>cABBBBBCD & F0DDFFIIKKLLNZPQQSSTTWWWWWXYZZ^O_i`i`j`k`z`{`egggggggiimmoooooqqrrss=u>ueggggGlllooo&{4{P{Q{$%'(),-CDFIJj))@ UVmH jU0JmH0J j0JUCJmH B*B* 5CJmH  5CJmH mH 6CJ$gggggggiimmoooooqqrrss=u>uwwyy%{&{'{({){*{+{,{-{.{/{0{1{2{3{4{P{Q{~~~~ %')*+,CGHIJg [>uwwyy%{&{'{({){*{+{,{-{.{/{0{1{2{3{4{P{Q{~~~~ & Fg )*+,h&`#$,GHIJ#0P$. A!"#$%`!MAY]9TaI&hY^?xMlUE_AQ-&5HĂn,&ǟHB!I\641 0nEjy̜3{罖 Ko{933oWA@x BXpMR X˪y;P4_AkhsT*oUMڒ;{GiWaY%_h\[Xu5t7XvآН˲=hz@l&{Rq0[:byb=%]'~ejЦ6>ئ~GFF/[%hwvhwvh7[yEZ<eO1:?Pü;$sӴ<&:(GSUŽΧ7b==^Z3vy a6[*EWS!{l}ׅΚcepėA2j.b ,GIkbO?uqbİnbnB/܆X fbPB] 7uҼLV$2w [E5>Ƅt8F=w>#LJzY'GSwa\.Mdxx(d*\&X<)L5רTKT?6[-%Gh)m$, Lcd*\&lz@2?IfL ? o $se6~O}o_oBO2Hfg{9I)WrX!{E{WuԕhVȽѴ;# Wt1{so(%$#1aaSx(EacjRPc*kQ|"]!B[uY$Hl&XتEC"?HuCjoTP`)Ӑh+Lҥ9 ˇF@5jcd{Y@fBtf}NkХZ^XlRPvkVHd`HY" i fY\L&,l|5L'3TQwW`m wV"s"s(%2w,s;1} e=YРgܮ[dne9= ~-_3KGzG2zgcйVJ:D2J:E'Zܵu"J֋䚕l'VQ$l~+ٌF-yV\hofu8Y+:IX=N(w'9]a5X9=:ؒ 察<0Di^YnB#$4v8*҃33ۃl{OG)s``4 6˙~e/\k 6/>≺U|l]9 |_vZpχnζBhTEVcWE.$첫\lڎ)u'NZ1bRߛ$~>΄t.J; (is# @v&{vffH6' د[oADqw' > dϸ];Ӝ]09t&dcA\qw>_"kpd=A|dCT ;Ȯ)l Lt3gALɮTBv9@q;G $[웎s6g~d?@cGv5vvgw>7X~Xn)ZwnXN~oѳ?ÇA ÅD!J8Ksq:$wqcpP?DZc86q>N~7mf~g9= z/"C'6-G'p.]Tˣ>J|b\2ɞxMHt1'#X3\͙Z Timָo%I_U?zt-Dd$< , C A? 2JMAY]9TaI&D`!MAY]9TaI&hY^?xMlUE_AQ-&5HĂn,&ǟHB!I\641 0nEjy̜3{罖 Ko{933oWA@x BXpMR X˪y;P4_AkhsT*oUMڒ;{GiWaY%_h\[Xu5t7XvآН˲=hz@l&{Rq0[:byb=%]'~ejЦ6>ئ~GFF/[%hwvhwvh7[yEZ<eO1:?Pü;$sӴ<&:(GSUŽΧ7b==^Z3vy a6[*EWS!{l}ׅΚcepėA2j.b ,GIkbO?uqbİnbnB/܆X fbPB] 7uҼLV$2w [E5>Ƅt8F=w>#LJzY'GSwa\.Mdxx(d*\&X<)L5רTKT?6[-%Gh)m$, Lcd*\&lz@2?IfL ? o $se6~O}o_oBO2Hfg{9I)WrX!{E{WuԕhVȽѴ;# Wt1{so(%$#1aaSx(EacjRPc*kQ|"]!B[uY$Hl&XتEC"?HuCjoTP`)Ӑh+Lҥ9 ˇF@5jcd{Y@fBtf}NkХZ^XlRPvkVHd`HY" i fY\L&,l|5L'3TQwW`m wV"s"s(%2w,s;1} e=YРgܮ[dne9= ~-_3KGzG2zgcйVJ:D2J:E'Zܵu"J֋䚕l'VQ$l~+ٌF-yV\hofu8Y+:IX=N(w'9]a5X9=:ؒ 察<0Di^YnB#$4v8*҃33ۃl{OG)s``4 6˙~e/\k 6/>≺U|l]9 |_vZpχnζBhTEVcWE.$첫\lڎ)u'NZ1bRߛ$~>΄t.J; (is# @v&{vffH6' د[oADqw' > dϸ];Ӝ]09t&dcA\qw>_"kpd=A|dCT ;Ȯ)l Lt3gALɮTBv9@q;G $[웎s6g~d?@cGv5vvgw>7X~Xn)ZwnXN~oѳ?ÇA ÅD!J8Ksq:$wqcpP?DZc86q>N~7mf~g9= z/"C'6-G'p.]Tˣ>J|b\2ɞxMHt1'#X3\͙Z Timָo%I_U?zt-  !"#$%&'()*+,-./0123456789:;<=>?@ABCDEFGHIJKLMNOPQRSTUVWXYZ[\]^_abcdefgjopqrstuvwxyz{|}~Root Entry  FlData `WordDocument KObjectPool\_1084172585 F\\Ole PRINTn'CompObje  FMicrosoft Excel-werkbladBiff8Excel.Sheet.89q Oh+'08@P\ tTFMEDIAFMEMicrosoft Excel@{ ՜.+,D՜.+,?- -    ''   Arial-Times New Roman-------Times New Roman----"System-'-   "--   !! "---'---  ---'---  !!- "-  $e!! - $e9e!$e!e99!e!] - "-  "- !e!leeEgege@e@eeeee>`e`e9e9---'---  !! - "-  $! e!! "- ] -$e9e!$e!e99!e!---'-- -  !! "- MM -  MMMM $hc -  $hh- ss -  ssss$hhh "-  M3 -   M3M3$c3f -   3f$- s&M -  s&Ms&M$ "-  f -   ff$oc -   $oo-  -  $ooo "-  f -   ff$yy6cy -   $yy-  -  $y66y "-  - MM -  MMMM$qccyq "-  - `` -  ````$iccqi "-  - M -  MM$b  ccib  "-  ff -   ffff$Z Z   cZ  -   $b b Z Z b -  -  $Z   b Z  "-  @ -   @@$R R )  cR -   $Z Z )R )R Z -` -  ``$R )  Z )R ) "-  -   $J J | cJ -  $R R J J R - - $J | |R J "- -  $B B   cB - $J J B B J - - $B   J B -- "- llEEee>>- 2 -0H 2 -1H 2 -2H 2 -3H 2 f-4H 2 -5H 2 ?-6H 2 -7H 2 -8H 2 -9H 2 10HH 2 _11HH 2 12HH-----'-- 31 Times New Roman-2 'ENo. of studies----'--  !!--'--  --'--  -- "- o o ] < "-- ^ ----'--- T 6---'--- T 6 "- -   M 2 Item 1%?*---'--- T 6---'--- T 6 "-3f - 3f   2 item 2%?*---'--- T 6---'--- T 6 "- -    2 0Item 3%?*---'--- T 6---'--- T 6 "- -   l8 2 Item 4%?*---'--- T 6---'--- T 6 "-ff - ff   2 Item 5%?*---'--- T 6---'--- T 6 "- -    2  Item 6%?*---'--- T 6---'--- T 6 "-f - f  \( 2 pItem 7%?*---'--- T 6---'--- T 6 "- -   x  2  Item 8%?*---'--- T 6---'--- T 6 "--      2  Item 9%?*---'-- - T 6---'-- - T 6 "--    L   2 ` Item 10%?**---'-- - T 6---'-- - T 6 "--     2  Item 11%?**---'--- T 6---'---  ---'---  -- -4-- jP/--- - --- -'   '  'ObjInfo Workbook#2SummaryInformation( DocumentSummaryInformation8 A\p Ba=< =ho<X@"1ZArial1ZArial1ZArial1ZArial1ZArial1ZArial1ZArial1ZArial1ZArial1ZArial16ZArial1.ZTimes New Roman1ZArial1.ZTimes New Roman7" "\ #,##0_-;" "\ #,##0\-A" "\ #,##0_-;[Red]" "\ #,##0\-C" "\ #,##0.00_-;" "\ #,##0.00\-M$" "\ #,##0.00_-;[Red]" "\ #,##0.00\-q*6_-" "\ * #,##0_-;_-" "\ * #,##0\-;_-" "\ * "-"_-;_-@_-,)'_-* #,##0_-;_-* #,##0\-;_-* "-"_-;_-@_-,>_-" "\ * #,##0.00_-;_-" "\ * #,##0.00\-;_-" "\ * "-"??_-;_-@_-4+/_-* #,##0.00_-;_-* #,##0.00\-;_-* "-"??_-;_-@_-"$"#,##0_);\("$"#,##0\)!"$"#,##0_);[Red]\("$"#,##0\)""$"#,##0.00_);\("$"#,##0.00\)'""$"#,##0.00_);[Red]\("$"#,##0.00\)72_("$"* #,##0_);_("$"* \(#,##0\);_("$"* "-"_);_(@_).)_(* #,##0_);_(* \(#,##0\);_(* "-"_);_(@_)?:_("$"* #,##0.00_);_("$"* \(#,##0.00\);_("$"* "-"??_);_(@_)61_(* #,##0.00_);_(* \(#,##0.00\);_(* "-"??_);_(@_) \p\o\i\n\t                + )  , * `Grafiek4 (Blad1 Q0Blad2 :1Blad3zr83  @@  7~~&#X7O +  FnH8bb{AHbWbHbI_H0b걃 ,1X1MA4"]~ Xs~s_Z Kl)M`L 6Fdٜ_ORI" dXX??3`  `  `  ` # ` +1(  ~  <a @AA?tc ]a0 $ 0Z0ZG3d523 M NM4 3Q Item 1QQQ3_4E4 3Q item 2QQQ3_4E4 3Q Item 3QQQ3_4E4 3Q Item 4QQQ3_4E4 3Q Item 5QQQ3_4E4 3Q Item 6QQQ3_4E4 3Q Item 7QQQ3_4E4 3Q Item 8QQQ3_4E4 3Q Item 9QQQ3_4E4 3Q Item 10QQQ 3_4E4 3Q Item 11QQQ 3_4E4D$% M 3O& Q4$% M 3O& Q4FA 3O 3 b#M! M! M 43*(@??!M& !  M! M! M! M NM4% d:M Z3O[P&Q  No. of studies'43" :dd *3O % M3O& Q423 M NM44444 ee&@$@@"@@? (@ $@e>   A  dMbP?_*+%"0??4U ( Ǭp0!>x p  6NMM? ]`0  AMLaserprinter? dXXL.HP LaserJet 4000 Series PCL 6E22xe.QwNO)"D/./`Eu݅ )M"H,jc'vvXx2X72̜9pM\n$Dd`CLAT؈j|Fvg18ꥯ̀t:0XZFJ#iVʹ^Ncz H <8me{"7aCE%{OaUC- fcvQBۡm#'w> ,1X1MA4"]~ Xs~s_Z Kl)M`L 6Fdٜ_ORI" dXX??3` ` `  3d23 M NM4 3Q Item 1QQQ3_4E4 3Q item 2QQQ3_4E4 3Q Item 3QQQ3_4E4 3Q Item 4QQQ3_4E4 3Q Item 5QQQ3_4E4 3Q Item 6QQQ3_4E4 3Q Item 7QQQ3_4E4 3Q Item 8QQQ3_4E4 3Q Item 9QQQ3_4E4 3Q Item 10QQQ 3_4E4 3Q Item 11QQQ 3_4E4D$% M 3O&Q4$% M 3O&Q4FAs 3Ot 3 b#M! M 43*,@??!M&!  M! M! M! M NM43" :, dd~ 3OI% M3O&Q423 M NM44444 ee&@$@@"@@? (@ $@e >@ A  dMbP?_*+%"??U>@ A  dMbP?_*+%"??U>@ P  PXt | Fontys Hogescholene1 Blad1Blad2Blad3 Grafiek4  Werkbladen Grafieken 6> _PID_GUIDAN{25DD674B-5E80-11D6-87F1-0002A57AC242}Oh+'0x1TableSummaryInformation(DocumentSummaryInformation8CompObjj   ( 4 @ LX`hp4sskulvikulvulvNormalRCzalen2zaMicrosoft Word 8.0@F#@ODE@n3 @n3 Lf՜.+,D՜.+,< hp  Fontys Hogescholen4}1 4 Title 6> _PID_GUIDAN{C3307EAC-7225-11D6-87F1-0002A581E2C0}  FMicrosoft Word Document MSWordDocWord.Document.89q [$@$NormalmH4@4 Heading 1$@&5mH 8@8 Heading 5 $$@&5mH <A@<Default Paragraph Font.B@. Body TextCJmH 2P@2 Body Text 26mH 8Q8 Body Text 35B*CJmH , @",Footer  p#&)@1& Page NumberJ|J| eJAJVH z#U+{++O,/Di`>u,JBDEFGIKLMNOQRSTUXYZa+gJCHPW !!:l,2$MAY]9TaI&%@(  @ N  S A ? B S  ? J|x#4 _1081930190 _1082215666 _1082479195 _1083997519 _1084171452 _1084172185 _1084172390 _1084172431 _1084172535 _1084172585C|C|C|C|C|C|C|C|C|C|K|@@@@@@@@@ @C|C|C|C|C|C|C|C|C|C|K|.3;@AH}`e q v /7LQ?Dnspypy,5AFDIR!W!&&N'S'|''''(&(((((K)P)6*;***++,,q/z/y115!5557778N;W;==BBDDEEEE GGGGNHXHgHsH PPVVxx|||(|+|,|F|H|K|. P V @G   $?D 2&&9'>'..00001122566677:B:Z>[>-I5INNRRSS__eejjlmnnnnrrttNu[uvv|||(|+|,|F|H|K|"A:\4.Discussion and conclusion.doc"A:\4.Discussion and conclusion.doc"A:\4.Discussion and conclusion.doc Gore Loro,C:\windows\TEMP\AutoHerstel-versie van 4.asd Gore Loro(A:\4.Discussion and conclusion final.doc Gore Loro(A:\4.Discussion and conclusion final.doc Gore Loro(A:\4.Discussion and conclusion final.doc Gore Loro(A:\4.Discussion and conclusion final.docICTheek(A:\4.Discussion and conclusion final.docRCzalen:C:\WIN95\Desktop\9.Chapter 4 Discussion and conclusion.doce+zF~A*n@<ZJc^; ~r N;$ d !\*L XwOGgN^psPG8h]W=,sutB=ܥ0 HtZ>[$oSFN BhQ<4G$B!V0r#;$/бsv$(,%as&:_ &b;l(V 9XH)h") {e-fk\.T3b/@o0T>/\/DE0+#5QN0<9| {054.Q(Vq14N։\x`4u,d4\5ze ;7paG7 5]v7`="8a:)u;4-^JzA.HHA l0BxwbB(nfgCHp9AyC$GLN GԌkIL>8\yM^'rOp+:aP #SaUT|UH)!@V [Xxj2YChXrZ2[[7^…#_踶[Qc*NOd){e~>V gj^)g*gCViʦb%oih,}Sj8bj*_xO mk,(sloJp,{plV,(!ufxr|=y..Wz\6(}oN\2^}8~(%~*..p.@ .......p.@ .......p.@ .......p.@ .......p.@ .......p.@ .......p.@ .....hho(->o(>o(.0o(..0o(... 88o( .... 88o( ..... `o( ...... `o(....... `o(........ hh>*CJo(.o(.808o(..808o(... 808o( .... o( ..... o( ...... `o(....... `o(........hho(-hho(. ?o(?o(.0o(..0o(... 88o( .... 88o( ..... `o( ...... `o(....... `o(........lo(lo(.(0(o(..T0To(... o( .... o( ..... `o( ...... `o(....... hho(........o(-hho(-hho(.hho(.hho(-?o(?o(.0o(..0o(... 88o( .... 88o( ..... `o( ...... `o(....... `o(........>o(>o(.0o(..0o(... 88o( .... 88o( ..... `o( ...... `o(....... `o(........XX5o(-o(-55o(-++o(.?o(?o(.0o(..0o(... (0(o( .... o( ..... o( ...... :`:o(....... P`Po(........hho(.hho(.XXo(-zo(zo(.0o(..0o(... 88o( .... 88o( ..... `o( ...... `o(....... `o(........XX5o(-hho(-hho(.\o(.0o(..0o(... 0o( .... 88o( ..... 88o( ...... `o(....... `o(........hho(hho(.0o(..88o(... 88o( .... `o( ..... `o( ...... o(....... ppo(........hh>*o(hh>*o(.0>*o(..0>*o(... 0>*o( .... 88>*o( ..... 88>*o( ...... `>*o(....... `>*o(........?o(?o(.0o(..0o(... 88o( .... 88o( ..... `o( ...... `o(....... o(........  o(-hho(-hho(-hho(.>o(>o(.0o(..0o(... 88o( .... 88o( ..... `o( ...... `o(....... `o(........   OJQJo(ffo(.o(.0o(..0o(... 0o( .... 88o( ..... 88o( ...... `o(....... `o(........hho(hho(.0o(..0o(... 0o( .... 88o( ..... 88o( ...... `o(....... `o(........zo(zo(.0o(..0o(... (0(o( .... o( ..... o( ...... :`:o(....... P`Po(........hho(hho(.0o(..0o(... 0o( .... 88o( ..... 88o( ...... `o(....... `o(........hho(.hhCJo(>o(>o(.0o(..0o(... 88o( .... 88o( ..... `o( ...... `o(....... `o(........hho()hho(hho(.0o(..0o(... 88o( .... 88o( ..... `o( ...... `o(....... `o(........hho(-o(.o(.808o(..808o(... 808o( .... o( ..... o( ...... `o(....... `o(........hho(.hho(.hh5o(-hho(-hho(hho(.0o(..88o(... 88o( .... `o( ..... `o( ...... o(....... ppo(........88o(-hho(.o(-o(.hho(.0o(..0o(... 0o( .... 88o( ..... 88o( ...... `o(....... `o(........>*o(\o(.0o(..88o(... 88o( .... `o( ..... `o( ...... o(....... ppo(........hho(.hho(hho(.0o(..88o(... 88o( .... `o( ..... `o( ...... o(....... ppo(........ hhB*CJo( B*CJo(. H0HB*CJo(.. B*CJo(... ((B*CJo( .... `B*CJo( ..... `B*CJo( ......  B*CJo(.......  P P B*CJo(........hho(o(.*0*o(..W0Wo(... 0o( .... o( ..... FFo( ...... `o(....... `o(........zo(zo(.0o(..0o(... 0o( .... 88o( ..... 88o( ...... `o(....... `o(........5OJQJo(hho(.?o(?o(.0o(..88o(... 88o( .... `o( ..... `o( ...... o(....... ppo(........hho(.hh>*o(.hho(.hho(. XXOJQJo(hho()hho(.hho(-hho(hho(.0o(..0o(... 88o( .... 88o( ..... `o( ...... `o(....... o(........0o(.0o(.0o(..88o(... 88o( .... `o( ..... `o( ...... o(....... ppo(........>o(>o(.0o(..0o(... 88o( .... 88o( ..... `o( ...... `o(....... `o(........hho(-88o(.hho(hho(.0o(..0o(... 0o( .... 88o( ..... 88o( ...... `o(....... `o(........0CJo(0CJo(.0CJo(..0CJo(... 88CJo( .... 88CJo( ..... `CJo( ...... `CJo(....... `CJo(........hho(-hho(hho(.0o(..0o(... 0o( .... 88o( ..... 88o( ...... `o(....... `o(........0o(0o(.0o(..88o(... 88o( .... `o( ..... `o( ...... o(....... ppo(........0o(.hho()hho()?o(?o(.0o(..0o(... 88o( .... 88o( ..... `o( ...... `o(....... o(........0o(0o(.0o(..0o(... 88o( .... 88o( ..... `o( ...... `o(....... o(........0o(0o(.0o(..0o(... 88o( .... 88o( ..... `o( ...... `o(....... o(........hho(-hho(-?o(?o(.0o(..0o(... 88o( .... 88o( ..... `o( ...... `o(....... `o(........o(-hho()gE0~r s&Jp%oi)g5QN0|U[$o!9AyC_ &cG$B!hSF$ #_W=rOo5]v7 | {0HAaG7:aPBB=b/l(OdOGkI ;$,{p(!uwbB$G[Qcbj2YXrZ̟tl0B#Su,d4gCVi2^}gC\.+`="8UT(}N^^; {e-G8u;FVq14 G542[[sl\x`4{e@<r|=y9XH)(tt ttttte ;7}Sj7^T>/,%A8\yM(%~0r#.Wza:O mkJzAO mkt!@V")[XO mktsuV g\*5sv$؟t Q@h OJQJo(lt0.t0.dt0.t0.\t0.آt0.Tt0.e@y~J|P@GTimes New Roman5Symbol3& Arial"0eePefLf40d}4kulvikRCzalen