Common use of Centres Clause in Contracts

Centres. All participants were seen in orthopaedic outpatient settings, at Gold Coast Hospital and Health Service, Australia. After the research assistant gained consent, all participants were independently examined the physiotherapist and the orthopaedic surgeon in variable order according to assessor availability. Patients were allocated 30 minutes with each assessor with the same information; the orthopaedic referral, the electronic hospital record, shoulder X-ray report, other results such as diagnostic imaging films and/or ACCEP ▇▇▇ ▇▇▇▇▇ CRIP T reports (MRI, CT, ultrasound scans) or pathology results that participants brought to the consultation. Each assessor performed a clinical assessment (history and physical examination) and completed a standardised assessment form for each participant, which was returned to the research assistant upon completion. Participants were blind to the profession of their assessors (i.e. they did not know which clinician was the physiotherapist and which was the surgeon), and were unaware of the clinical decisions of the assessors until all assessments had been completed. Assessors consulted with participants independently, in separate rooms and were blind to each other’s findings. Participant characteristics collected by the research assistant prior to assessment included demographics, symptom duration, the Shoulder Pain and Disability Index (SPADI) which is a reliable and valid self-rating tool for people with shoulder pain (Angst et al. 2011; ▇▇▇▇▇ et al. 1991); the worst shoulder pain severity over the past three days via a 100 mm visual analogue scale (VAS; 0 = no pain, 100 = worst pain imaginable); and the European Quality of Life five dimensions, five levels (EQ-5D-5L), converted to Australian values (▇▇▇▇▇▇, ▇▇▇▇▇▇, and ▇▇▇▇▇ 2013). Assessors recorded their findings and clinical decisions on the paper assessment form. The research assistant was responsible for entry of data into the electronic database and coding of free text into categories for analysis. ACCEP ▇▇▇ ▇▇▇▇▇ CRIP T Primary outcomes investigated management and subacromial corticosteroid injection decisions made by each assessor. Management was recorded in two ways: dichotomous response (yes / no) to the question “is it appropriate to have initial non-surgical care?”, and by free text “proposed management plan” responses. Subacromial corticosteroid injection decisions were investigated with dichotomous (yes / no) assessor responses to three questions focusing on different considerations within an injection decision: safety, “is it safe to offer injection today?”; whether there is symptomatic indication for the injection, “do clinical findings support the use of subacromial injection?”; and whether injection is the immediate treatment priority, “would you provide subacromial injection today?”. Assessors also recorded reasons for not offering same-day injection via free text and under categories of safety, consent, other priority, need for prior investigation, or other. Secondary outcomes were diagnoses and radiology or pathology investigations recorded in free text. Diagnoses were assigned to one of eight categories by the research assistant (if multiple contributory diagnoses were recorded, the first or primary diagnosis was used): 1. subacromial impingement syndrome (SAIS, SIS, bursitis), 2. glenohumeral osteoarthritis, 3. acromioclavicular osteoarthritis, 4. adhesive capsulitis, 5. rotator cuff tear (distinguished by it being listed above any other factors or requiring surgical repair), 6. long head of biceps symptoms, 7. instability and / or labral pathology, and 8. pain from non-shoulder origin (including cervical referred, systemic inflammatory disorders, neurological or sensory disorders). ACCEP ▇▇▇ ▇▇▇▇▇ CRIP T Sample size calculations for the randomised controlled trial within the published protocol (▇▇▇▇▇ et al. 2014), assumed α = 0.05 and power = 0.8 (80%), with a sample of 64 participants required for the RCT. We anticipated that 25% of all participants assessed, would enter the RCT. As all assessed participants would be eligible for the present agreement study, 256 were expected and 274 participants were ultimately recruited. This number exceeds the largest number of subjects (n=100) previously reported by an inter-rater orthopaedic and physiotherapy agreement study (▇▇▇▇▇▇▇ et al. 2013). It also exceeds the number required to detect Kappa values over 0.6 with 90% power, assuming a null of Kappa = 0.4 (Sim and ▇▇▇▇▇▇ 2005), and the number required for agreement proportions of 30% to 100%, with a relative error of 20% (Gwet 2010). Data were analysed using SPSS software version 22 (IBM, Chicago USA), AgreeStat 2013.4, and AgreeStat 2015.5 for Excel (Advanced Analytics; Gaithersburg, MD, USA). Accuracy of the free text coding was assessed by two investigators (DM, LB), and at the completion of data entry the accuracy of the data entries ensured through comparison of the hard copy data and the data on the electronic database for 30 participants (every tenth participant), by one investigator (DM). Descriptive statistics were used to describe baseline characteristics. Four agreement coefficients were calculated for primary and secondary outcomes, to assess the level of agreement (inter-rater reliability) between the physiotherapist and the orthopaedic surgeon: ▇▇▇▇▇’▇ Kappa (▇▇▇ and ▇▇▇▇▇▇ 2005), prevalence and bias adjusted Kappa ACCEP ▇▇▇ ▇▇▇▇▇ CRIP T (▇▇▇▇▇) (▇▇▇ and ▇▇▇▇▇▇ 2005), ▇▇▇▇’s first order agreement coefficient (AC1) (Gwet 2012), and percentage agreement. For free text “proposed management plan” responses, frequencies were calculated and presented descriptively. Descriptive analysis of disagreements was also undertaken. Kappa is commonly used for inter-rater agreement calculations (Gwet 2012; ▇▇▇ and ▇▇▇▇▇▇ 2005), yet its paradoxes are known to cause low values in the presence of high observed agreement (▇▇▇▇▇▇▇▇▇ and ▇▇▇▇▇▇▇▇▇ 1990). To overcome this we used PABAK and ▇▇▇▇’s AC1 which both provide different alternative calculations for the agreement by chance (Gwet 2012). PABAK removes bias by bringing the expected agreement to an average value of 0.5 and should be interpreted alongside ▇▇▇▇▇’▇ Kappa (▇▇▇ and ▇▇▇▇▇▇ 2005) and individual cell data (▇▇▇▇ et al. 2009). Gwet’s AC1 removes the agreement by chance (Gwet 2012) and has been recommended for its stability in the presence of high inter-rater agreement (▇▇▇▇▇▇▇▇▇▇▇ et al. 2013). As previous shoulder inter-rater agreement studies used Kappa and PABAK (▇▇▇▇▇▇▇ et al. 2013), or Kappa with percentage agreement (▇▇▇▇▇ and ▇▇▇▇▇▇ 2008), we have presented the four measures, to both aid comparison with previous research and to demonstrate the consistency of our findings with the different measures. We applied the following established scale (▇▇▇▇▇▇ and ▇▇▇▇ 1977) to the interpretation of the magnitude of Kappa and AC1 values: <0.2 poor, 0.21 - 0.4 fair, 0.41 - 0.6 moderate, 0.61 - 0.8 substantial, 0.81 – 1 near perfect agreement. ACCEP ▇▇▇ ▇▇▇▇▇ CRIP T 988 orthopaedic referrals were screened for eligibility by the research assistant according to the published protocol (Marks et al. 2014). Of 305 attending between January 2013 and June 2014, 278 met the inclusion criteria. Three did not complete assessment at completion of the RCT protocol, and one participant with bilateral shoulder pain was excluded as the assessors each examined different shoulders. This left 274 participants who were examined by the physiotherapist and the orthopaedic surgeon, their characteristics are described in Table 1.

Centres. All participants were seen in orthopaedic outpatient settings, at Gold Coast Hospital and Health Service, AustraliaXXXX. Intervention (clinical assessments) After the research assistant gained consent, all participants were independently examined the physiotherapist and the orthopaedic surgeon in variable order according to assessor availability. Patients were allocated 30 minutes with each assessor with the same information; the orthopaedic referral, the electronic hospital record, shoulder X-ray X−ray report, other results such as diagnostic imaging films and/or ACCEP ▇▇▇ ▇▇▇▇▇ CRIP T reports (MRI, CT, ultrasound scans) or pathology results that participants brought to the consultation. Each assessor performed a clinical assessment (history and physical examination) and completed a standardised assessment form for each participant, which was returned to the research assistant upon completion. Participants were blind to the profession of their assessors (i.e. they did not know which clinician was the physiotherapist and which was the surgeon), and were unaware of the clinical decisions of the assessors until all assessments had been completed. Assessors consulted with participants independently, in separate rooms and were blind to each other’s findings. Outcome measures Participant characteristics collected by the research assistant prior to assessment included demographics, symptom duration, the Shoulder Pain and Disability Index (SPADI) which is a reliable and valid self-rating self−rating tool for people with shoulder pain (Angst et al. 2011; ▇▇▇▇▇ et al. 1991); the worst shoulder pain severity over the past three days via a 100 mm visual analogue scale (VAS; 0 = no pain, 100 = worst pain imaginable); and the European Quality of Life five dimensions, five levels (EQ-5D-5LEQ−5D−5L), converted to Australian values (▇▇▇▇▇▇Norman, ▇▇▇▇▇▇Cronin, and ▇▇▇▇▇ 2013). Assessors recorded their findings and clinical decisions on the paper assessment form. The research assistant was responsible for entry of data into the electronic database and coding of free text into categories for analysis. ACCEP ▇▇▇ ▇▇▇▇▇ CRIP T Primary outcomes investigated management and subacromial corticosteroid injection decisions made by each assessor. Management was recorded in two ways: dichotomous response (yes / no) ふyes っ noぶ to the question “is ケuestion さis it appropriate to have ha┗e initial non-surgical care?”non−surgical careいざ, and by free text “proposed te┝t さproposed management plan” planざ responses. Subacromial corticosteroid injection decisions were investigated with dichotomous (yes / no) assessor responses to three questions focusing on different considerations within an injection decision: safety, “is さis it safe to offer injection today?”; inﾃection todayいざき whether there is symptomatic indication for the injectioninﾃection, “do さdo clinical findings support the use of subacromial injection?”; injection?ざき and whether injection is the immediate treatment priority, “would さwould you provide pro┗ide subacromial injection today?”inﾃection todayいざ. Assessors also recorded reasons for not offering same-day same−day injection via free text and under categories of safety, consent, other priority, need for prior investigation, or other. Secondary outcomes were diagnoses and radiology or pathology investigations recorded in free text. Diagnoses were assigned to one of eight categories by the research assistant (if multiple contributory diagnoses were recorded, the first or primary diagnosis was used): 1. subacromial impingement syndrome (SAIS, SIS, bursitis), 2. glenohumeral osteoarthritis, 3. acromioclavicular osteoarthritis, 4. adhesive capsulitis, 5. rotator cuff tear (distinguished by it being listed above any other factors or requiring surgical repair), 6. long head of biceps symptoms, 7. instability and / or labral pathology, and 8. pain from non-shoulder non−shoulder origin (including cervical referred, systemic inflammatory disorders, neurological or sensory disorders). ACCEP ▇▇▇ ▇▇▇▇▇ CRIP T Sample size Sample size calculations for the randomised controlled trial within the published protocol (▇▇▇▇▇ XXXX et al. 2014), assumed α a = 0.05 and power β = 0.8 (80%)0.2, with a sample of 64 participants required for the RCT. We anticipated that 25% of all participants assessed, would enter the RCT. As all assessed participants would be eligible for the present agreement study, 256 were expected and 274 participants were ultimately recruited. This number exceeds the largest number of subjects (n=100) previously reported by an inter-rater inter−rater orthopaedic and physiotherapy agreement study (▇▇▇▇▇▇▇ et al. 2013). It also exceeds the number required to detect Kappa values over 0.6 with 90% power, assuming a null of Kappa = 0.4 (Sim and ▇▇▇▇▇▇ 2005), and the number required for agreement proportions of 30% to 100%, with a relative error of 20% (Gwet 2010). Data analysis Data were analysed using SPSS software version 22 (IBM, Chicago USA), AgreeStat 2013.4, and AgreeStat 2015.5 for Excel (Advanced Analytics; Gaithersburg, MD, USA). Accuracy of the free text coding was assessed by two investigators (DMXX, LBXX), and at the completion of data entry the accuracy of the data entries ensured through comparison of the hard copy data and the data on the electronic database for 30 participants (every tenth participant), by one investigator (DMXX). Descriptive statistics were used to describe baseline characteristics. Four agreement coefficients were calculated for primary and secondary outcomes, to assess the level of agreement (inter-rater inter−rater reliability) between the physiotherapist and the orthopaedic surgeon: ▇▇▇▇▇’▇ Kappa (▇▇▇ and ▇▇▇▇▇▇ 2005), prevalence and bias adjusted Kappa ACCEP ▇▇▇ ▇▇▇▇▇ CRIP T (▇▇▇▇▇) (▇▇▇ and ▇▇▇▇▇▇ 2005▇▇▇▇), ▇▇▇▇’s first order agreement coefficient (AC1) (Gwet 2012), and percentage agreement. For free text “proposed management plan” responses, frequencies were calculated and presented descriptively. Descriptive analysis of disagreements was also undertaken. Kappa is commonly used for inter-rater agreement calculations (Gwet 2012; ▇▇▇ and ▇▇▇▇▇▇ 2005), yet its paradoxes are known to cause low values in the presence of high observed agreement (▇▇▇▇▇▇▇▇▇ and ▇▇▇▇▇▇▇▇▇ 1990). To overcome this we used PABAK and ▇▇▇▇’s AC1 which both provide different alternative calculations for the agreement by chance (Gwet 2012). PABAK removes bias by bringing the expected agreement to an average value of 0.5 and should be interpreted alongside ▇▇▇▇▇’▇ Kappa (▇▇▇ and ▇▇▇▇▇▇ 2005) and individual cell data (▇▇▇▇ et al. 2009). Gwet’s AC1 removes the agreement by chance (Gwet 2012) and has been recommended for its stability in the presence of high inter-rater agreement (▇▇▇▇▇▇▇▇▇▇▇ et al. 2013). As previous shoulder inter-rater agreement studies used Kappa and PABAK (▇▇▇▇▇▇▇ et al. 2013), or Kappa with percentage agreement (▇▇▇▇▇ and ▇▇▇▇▇▇ 2008), we have presented the four measures, to both aid comparison with previous research and to demonstrate the consistency of our findings with the different measures. We applied the following established scale (▇▇▇▇▇▇ and ▇▇▇▇ 1977) to the interpretation of the magnitude of Kappa and AC1 values: <0.2 poor, 0.21 - 0.4 fair, 0.41 - 0.6 moderate, 0.61 - 0.8 substantial, 0.81 – 1 near perfect agreement. ACCEP ▇▇▇ ▇▇▇▇▇ CRIP T 988 orthopaedic referrals were screened for eligibility by the research assistant according to the published protocol (Marks et al. 2014). Of 305 attending between January 2013 and June 2014, 278 met the inclusion criteria. Three did not complete assessment at completion of the RCT protocol, and one participant with bilateral shoulder pain was excluded as the assessors each examined different shoulders. This left 274 participants who were examined by the physiotherapist and the orthopaedic surgeon, their characteristics are described in Table 1.

Centres. All participants were seen in orthopaedic outpatient settings, at Gold Coast Hospital and Health Service, AustraliaXXXX. Intervention (clinical assessments) After the research assistant gained consent, all participants were independently examined the physiotherapist and the orthopaedic surgeon in variable order according to assessor availability. Patients were allocated 30 minutes with each assessor with the same information; the orthopaedic referral, the electronic hospital record, shoulder X-ray X−ray report, other results such as diagnostic imaging films and/or ACCEP ▇▇▇ ▇▇▇▇▇ CRIP T reports (MRI, CT, ultrasound scans) or pathology results that participants brought to the consultation. Each assessor performed a clinical assessment (history and physical examination) and completed a standardised assessment form for each participant, which was returned to the research assistant upon completion. Participants were blind to the profession of their assessors (i.e. they did not know which clinician was the physiotherapist and which was the surgeon), and were unaware of the clinical decisions of the assessors until all assessments had been completed. Assessors consulted with participants independently, in separate rooms and were blind to each other’s findings. Outcome measures Participant characteristics collected by the research assistant prior to assessment included demographics, symptom duration, the Shoulder Pain and Disability Index (SPADI) which is a reliable and valid self-rating self−rating tool for people with shoulder pain (Angst et al. 2011; ▇▇▇▇▇ et al. 1991); the worst shoulder pain severity over the past three days via a 100 mm visual analogue scale (VAS; 0 = no pain, 100 = worst pain imaginable); and the European Quality of Life five dimensions, five levels (EQ-5D-5LEQ−5D−5L), converted to Australian values (▇▇▇▇▇▇Norman, ▇▇▇▇▇▇Cronin, and ▇▇▇▇▇ 2013). Assessors recorded their findings and clinical decisions on the paper assessment form. The research assistant was responsible for entry of data into the electronic database and coding of free text into categories for analysis. ACCEP ▇▇▇ ▇▇▇▇▇ CRIP T Primary outcomes investigated management and subacromial corticosteroid injection decisions made by each assessor. Management was recorded in two ways: dichotomous response (yes / no) ふyes っ noぶ to the question “is ケuestion さis it appropriate to have ha┗e initial non-surgical care?”non−surgical careいざ, and by free text “proposed te┝t さproposed management plan” planざ responses. Subacromial corticosteroid injection decisions were investigated with dichotomous (yes / no) assessor responses to three questions focusing on different considerations within an injection decision: safety, “is さis it safe to offer injection today?”; inﾃection todayいざき whether there is symptomatic indication for the injectioninﾃection, “do さdo clinical findings support the use of subacromial injection?”; injection?ざき and whether injection is the immediate treatment priority, “would さwould you provide pro┗ide subacromial injection today?”inﾃection todayいざ. Assessors also recorded reasons for not offering same-day same−day injection via free text and under categories of safety, consent, other priority, need for prior investigation, or other. Secondary outcomes were diagnoses and radiology or pathology investigations recorded in free text. Diagnoses were assigned to one of eight categories by the research assistant (if multiple contributory diagnoses were recorded, the first or primary diagnosis was used): 1. subacromial impingement syndrome (SAIS, SIS, bursitis), 2. glenohumeral osteoarthritis, 3. acromioclavicular osteoarthritis, 4. adhesive capsulitis, 5. rotator cuff tear (distinguished by it being listed above any other factors or requiring surgical repair), 6. long head of biceps symptoms, 7. instability and / or labral pathology, and 8. pain from non-shoulder non−shoulder origin (including cervical referred, systemic inflammatory disorders, neurological or sensory disorders). ACCEP ▇▇▇ ▇▇▇▇▇ CRIP T Sample size Sample size calculations for the randomised controlled trial within the published protocol (▇▇▇▇▇ XXXX et al. 2014), assumed α a = 0.05 and power β = 0.8 (80%)0.2, with a sample of 64 participants required for the RCT. We anticipated that 25% of all participants assessed, would enter the RCT. As all assessed participants would be eligible for the present agreement study, 256 were expected and 274 participants were ultimately recruited. This number exceeds the largest number of subjects (n=100) previously reported by an inter-rater inter−rater orthopaedic and physiotherapy agreement study (▇▇▇▇▇▇▇ et al. 2013). It also exceeds the number required to detect Kappa values over 0.6 with 90% power, assuming a null of Kappa = 0.4 (Sim and ▇▇▇▇▇▇ 2005), and the number required for agreement proportions of 30% to 100%, with a relative error of 20% (Gwet 2010). Data analysis Data were analysed using SPSS software version 22 (IBM, Chicago USA), AgreeStat 2013.4, and AgreeStat 2015.5 for Excel (Advanced Analytics; Gaithersburg, MD, USA). Accuracy of the free text coding was assessed by two investigators (DMXX, LBXX), and at the completion of data entry the accuracy of the data entries ensured through comparison of the hard copy data and the data on the electronic database for 30 participants (every tenth participant), by one investigator (DMXX). Descriptive statistics were used to describe baseline characteristics. Four agreement coefficients were calculated for primary and secondary outcomes, to assess the level of agreement (inter-rater inter−rater reliability) between the physiotherapist and the orthopaedic surgeon: ▇▇▇▇▇’▇ Kappa (▇▇▇ and ▇▇▇▇▇▇ 2005), prevalence and bias adjusted Kappa ACCEP ▇▇▇ ▇▇▇▇▇ CRIP T (▇▇▇▇▇) (▇▇▇ and ▇▇▇▇▇▇ 2005), ▇▇▇▇’s first order agreement coefficient (AC1) (Gwet 2012), and percentage agreement. For free text “proposed management plan” responses, frequencies were calculated and presented descriptively. Descriptive analysis of disagreements was also undertaken. Kappa is commonly used for inter-rater agreement calculations (Gwet 2012; ▇▇▇ and ▇▇▇▇▇▇ 2005), yet its paradoxes are known to cause low values in the presence of high observed agreement (▇▇▇▇▇▇▇▇▇ and ▇▇▇▇▇▇▇▇▇ 1990). To overcome this we used PABAK and ▇▇▇▇’s AC1 which both provide different alternative calculations for the agreement by chance (Gwet 2012). PABAK removes bias by bringing the expected agreement to an average value of 0.5 and should be interpreted alongside ▇▇▇▇▇’▇ Kappa (▇▇▇ and ▇▇▇▇▇▇ 2005) and individual cell data (▇▇▇▇ et al. 2009). Gwet’s AC1 removes the agreement by chance (Gwet 2012) and has been recommended for its stability in the presence of high inter-rater agreement (▇▇▇▇▇▇▇▇▇▇▇ et al. 2013). As previous shoulder inter-rater agreement studies used Kappa and PABAK (▇▇▇▇▇▇▇ et al. 2013), or Kappa with percentage agreement (▇▇▇▇▇ and ▇▇▇▇▇▇ 2008), we have presented the four measures, to both aid comparison with previous research and to demonstrate the consistency of our findings with the different measures. We applied the following established scale (▇▇▇▇▇▇ and ▇▇▇▇ 1977) to the interpretation of the magnitude of Kappa and AC1 values: <0.2 poor, 0.21 - 0.4 fair, 0.41 - 0.6 moderate, 0.61 - 0.8 substantial, 0.81 – 1 near perfect agreement. ACCEP ▇▇▇ ▇▇▇▇▇ CRIP T 988 orthopaedic referrals were screened for eligibility by the research assistant according to the published protocol (Marks et al. 2014). Of 305 attending between January 2013 and June 2014, 278 met the inclusion criteria. Three did not complete assessment at completion of the RCT protocol, and one participant with bilateral shoulder pain was excluded as the assessors each examined different shoulders. This left 274 participants who were examined by the physiotherapist and the orthopaedic surgeon, their characteristics are described in Table 1.