Transcript Teal
“Show Me How to Get Past MCQs: Emerging Opportunities in Measurement ” Carol O’Byrne, PEBC Karen S. Flint and Jaime Walla, AMP Drs. Frank Hideg, Paul Townsend, & Mark Christensen, NBCE Alison Cooper, CAPR Lila Quero-Munoz, Consultant Presented at the 2004 CLEAR Annual Conference September 30 – October 2 Kansas City, Missouri Goals 1. 2. 3. 4. 5. 6. Gain an overview of performance assessment Observe and try out electronic & standardized patient simulations Consider exam development, implementation and administration issues Consider validity questions & research needs Create computer-administered & standardized patient simulations with scoring rubrics Set passing standards Presented at the 2004 CLEAR Annual Conference September 30 – October 2 Kansas City, Missouri Part 1 - Presentations Introduction to performance assessment – Purposes and objectives – Models – Issues, successes and challenges 15-minute presentations – Four models, including their unique aspects with two participatory demonstrations – Developmental and ongoing validity issues and research studies Presented at the 2004 CLEAR Annual Conference September 30 – October 2 Kansas City, Missouri Part 2 - Break-out Sessions 1. Identify steps in development and implementation of a new performance assessment and develop a new station 2. Create a new electronic simulation and set passing standards 3. Create a new standardized patient simulation and scoring rubrics 4. Participate in a standard setting exercise using the ‘Competence Standard Setting Method’ and all the while, ask the ‘hard questions’ Presented at the 2004 CLEAR Annual Conference September 30 – October 2 Kansas City, Missouri Performance Assessment WHY? To assess important problem solving, critical thinking, communications, hands-on and other complex skills that: – Impact clients' safety and welfare if not performed adequately and – Are difficult to assess in a multiple choice question format Presented at the 2004 CLEAR Annual Conference September 30 – October 2 Kansas City, Missouri HOW? • ‘Pot luck’ direct observation (e.g., medical rounds, clerkships and internships) • Semi-structured assessments (e.g. orals and Patient Management Problems) • Objective, Structured Clinical Examinations (OSCEs) (combining standardized client interactions with other formats) • Other standardized simulations (e.g., airline pilots' simulators) • Electronic simulations (e.g., real estate, respiratory care, architecture) Presented at the 2004 CLEAR Annual Conference September 30 – October 2 Kansas City, Missouri Does it really work? Links in the Chain of Evidence to Support the Validity of Examination Results: ℴ ℴ ℴ ℴ ℴ ℴ ℴ ℴ Job Analysis Test Specifications Item Writing Examination Construction Standard Setting Test Administration Scoring Reporting Test Results Presented at the 2004 CLEAR Annual Conference September 30 – October 2 Kansas City, Missouri PEBC Qualifying Examination • Based on national competencies • Two parts: MCE & OSCE • Must pass both to be eligible for pharmacist licensure in Canada • Offered spring and fall in multiple locations • 1400+ candidates/year • $1350 CDN • 15-station OSCE – 12 client interactions (SP or SHP) + 3 non-client stations – 7 minute stations • One expert examiner • Checklist to document performance • Holistic ratings to score exam • Standard Setting • Reports – results and feedback Presented at the 2004 CLEAR Annual Conference September 30 – October 2 Kansas City, Missouri Competencies Assessed by PEBC’s MCE and OSCE COMPETENCIES MCE OSCE % % 56.5 29 2: Assume ethical, legal and professional responsibilities 8.5 9 3: Access, retrieve, evaluate and disseminate relevant information 6.5 5 4: Communicate and educate effectively 1.5 43 23 9 4 5 1. Practise pharmaceutical care 5: Manage drug distribution 6: Apply practice management knowledge and skills Presented at the 2004 CLEAR Annual Conference September 30 – October 2 Kansas City, Missouri Comparing PEBC’s OSCE (PS04) and MCE (QS04) Scores Paired Samples Statistics Pair 1 PS04 QS04 Mean 473.7354 481.2897 N 711 711 Std. Deviation 95.14097 96.62295 Std. Error Mean 3.56807 3.62364 Paired Samples Correlations N Pair 1 PS04 & QS04 711 Correlation .607 Presented at the 2004 CLEAR Annual Conference September 30 – October 2 Kansas City, Missouri Sig. .000 Comparing PEBC’s OSCE and MCE scores Paired Samples Statistics OriginS04 1 Pair 1 2 Pair 1 USA Pair 1 PS04 QS04 PS04 QS04 PS04 QS04 Mean 500.0402 502.4589 353.7158 387.7982 445.4000 448.9459 N 560 560 114 114 37 37 Std. Deviation 68.96394 85.94416 109.35265 88.43879 99.66310 95.69075 Std. Error Mean 2.91426 3.63180 10.24181 8.28305 16.38451 15.73146 Paired Samples Correlations OriginS04 1 2 USA N Pair 1 Pair 1 Pair 1 PS04 & QS04 PS04 & QS04 PS04 & QS04 560 114 37 Correlation .445 .571 .688 Presented at the 2004 CLEAR Annual Conference September 30 – October 2 Kansas City, Missouri Sig. .000 .000 .000 Holistic Rating Scales COMMUNICATION Skills (1) 100 80 Percent – Rapport – Organization – Verbal and nonverbal expression Holistic Scale Means Problem-solving OUTCOME (2) 0 1 2 Holistic Scale 3 Analytical Score Means Overall PERFORMANCE (3) 80 Percent Comm & Outcome Thoroughness (checklist) Accuracy (misinformation) Risk 40 20 – Information processing – Decision making – Follow-up – – – – 60 60 40 20 0 1 2 3 4 5 Competencies Presented at the 2004 CLEAR Annual Conference September 30 – October 2 Kansas City, Missouri 6 PEBC QUALIFYING EXAMINATION - PART II (OSCE) EXAMINATION FEEDBACK REPORT Candidate ID Table 1 Status: Fail RATINGS Communication Your average 3.10 Group average 3.67 Rating 4=Acceptable 3=Marginally acceptable 2=Marginally unacceptable 1=Unacceptable Outcomes Your average 2.33 Group average 2.90 Rating # of Stations (of 15) 4=Problem solved 2 3=Solved marginally 5 2=Uncertain/marginally unsolved 4 1=Problem unsolved 4 Overall Performance Your average 2.51 Group average 3.01 Rating 4=Acceptable 3=Marginally acceptable 2=Marginally unacceptable 1=Unacceptable Table 2 # of Stations (of 15) 3 5 4 3 MISINFORMATION AND RISK/INEFFECTIVE THERAPY Misinformation Risk or ineffective therapy Table 3 # of Stations (of 12) 4 6 1 1 # of Stations/Instances (of 15) 3 2 Group Average 1.31 0.63 COMPETENCY SCORES Competency Your Score (%) 1 Practise pharmaceutical care 47 2 Assume ethical, legal and professional responsibilities 39 3 Access, retrieve, evaluate, disseminate relevant information 44 4 Communicate and educate effectively 35 5 Manage drug distribution 80 6 Apply practice management knowledge and skills 50 Group Average (%) 59 56 47 53 74 63 Presented at the 2004 CLEAR Annual Conference September 30 – October 2 Kansas City, Missouri Validity – an ascent from Practice Analysis to Test Results Job/practice analysis – Who/what contexts? – How? Test specifications & sampling – Which competencies? – Which tasks/scenarios? – Other parameters? Item writing and review – Who and how? Scoring – Analytic (checklists) &/or holistic (scales)? Presented at the 2004 CLEAR Annual Conference September 30 – October 2 Kansas City, Missouri Validity – an ascent from Practice Analysis to Test Results Detect and minimize unwanted variability, e.g.: – Items/tasks – does the mix matter? – Practice effect – how can we avoid it? – Presentation/administration – what is the impact of different SPs, computers, materials/equipment? – Scores – how do we know how accurate and dependable they are? What can we do to improve accuracy? Set Defensible Pass-fail Standards – How should we do this when different standard setting methods > different standards? – How do we know if the standard is appropriate? Report Results – Are they clear? Interpreted correctly? – Are they defensible? Presented at the 2004 CLEAR Annual Conference September 30 – October 2 Kansas City, Missouri Validity – flying high Evidence – Strong links from job analysis to interpretation of test results – Relates to performance in training & other tests Reliable, generalizable & dependable – Scores – Pass-fail standards & outcomes Feasible – Large & small scale programs – Economic, human, physical, technological resources Ongoing Research Presented at the 2004 CLEAR Annual Conference September 30 – October 2 Kansas City, Missouri Wild Life Candidate diversity – Language – Training – Format familiarity, e.g. computer skills – Accommodations Logistics – Technological requirements – Replications (fatigue, attention span) Security Presented at the 2004 CLEAR Annual Conference September 30 – October 2 Kansas City, Missouri “Computer-Based Simulations” Karen S. Flint Director, Internal Development & Systems Integration Applied Measurement Professionals, Inc. Presented at the 2004 CLEAR Annual Conference September 30 – October 2 Kansas City, Missouri Evolution of Simulation Exam Format • AMP’s parent company, NBRC, provided oral exams from 1961 to 1978 • Alternative sought due to: – Limited number of candidates that could be tested each administration – Cost to candidates who had to travel to location – Concern about potential oral examiner bias Presented at the 2004 CLEAR Annual Conference September 30 – October 2 Kansas City, Missouri Evolution of Simulation Exam Format • Printed simulation exam format introduced in 1978 using latent image technology • Latent image format used by NBRC from 1978 to 1999 • NBRC decision to convert all exams to computer-based testing • Proprietary software developed by AMP to administer simulation exams in comparable format via computer – introduced in 2000 • Both latent image test booklets & computerized format being used Presented at the 2004 CLEAR Annual Conference September 30 – October 2 Kansas City, Missouri How Simulation Exams Differ from MCQs • Provides accurate assessment of higher order thinking related to a content area of interest (testing more than just recall) • Challenge test takers beyond complexity of MCQs • Simulation problems allow test takers to assess their skills against test content drawn from realistic situations or clinical events Presented at the 2004 CLEAR Annual Conference September 30 – October 2 Kansas City, Missouri Sample relationship between multiple-choice and simulation scores assessing similar content 150 145 Passers 140 135 130 125 120 115 110 105 100 Failers 95 90 15 20 25 30 35 40 45 Scaled Simulation Score Note: n = 68, r = .48, p <.001 Presented at the 2004 CLEAR Annual Conference September 30 – October 2 Kansas City, Missouri 50 Simulation Utility • Continuing competency examinations • Self-assessment/practice examinations • High-stakes examinations – Psychometric characteristics comparable to other assessment methodologies – That is, good reliability and validity Presented at the 2004 CLEAR Annual Conference September 30 – October 2 Kansas City, Missouri Professions Using This Simulation Format • • • • • Advanced-Level Respiratory Therapists Advanced-Level Dietitians Lighting Design Professionals Orthotist/Prosthetist Professionals Health System Case Management Professionals (beginning 2005) • Real Estate Professionals • Candidate fees range from $200 to $525 for full-length certification/licensure simulation exam Presented at the 2004 CLEAR Annual Conference September 30 – October 2 Kansas City, Missouri Structure of Simulations • Opening Scenario • Information Gathering (IG) Sections • Decision Making (DM) Sections – Single or multiple DM • All choices are weighted (+3 to –3) • Passing scores relate to judgment of content experts on ‘minimal competence’ Presented at the 2004 CLEAR Annual Conference September 30 – October 2 Kansas City, Missouri Simulation Development (Graphic depiction of path through a simulation problem) Presented at the 2004 CLEAR Annual Conference September 30 – October 2 Kansas City, Missouri IG Section Details • IG section – A section in which test takers choose information that will best help them understand a presenting problem or situation – Facilitative options may receive scores of +3, +2, or +1 – Uninformative, wasteful, unnecessarily invasive, or potentially illegal options may receive scores of –1, – 2, or –3 – Test takers who select undesirable options accumulate negative section points Presented at the 2004 CLEAR Annual Conference September 30 – October 2 Kansas City, Missouri IG Section Details • IG Section Minimum Pass Level (MPL) – Among all options with positive scores in a section, some should be designated as REQUIRED for minimally competent practice – The sum of points for all REQUIRED options in a section equals MPL Presented at the 2004 CLEAR Annual Conference September 30 – October 2 Kansas City, Missouri DM Section Details • DM section – A section of typically 4-6 options in which the test taker must make a decision about how to handle the presenting situation – Facilitative options may receive scores of +3, +2, or +1 – Harmful or potentially illegal options may receive scores of –1, –2, or –3 – Test takers who select undesirable options accumulate negative section points and are directed to select another option Presented at the 2004 CLEAR Annual Conference September 30 – October 2 Kansas City, Missouri DM Section Details • DM Section Minimum Pass Level (MPL) – May contain two correct choices, but one must be designated as REQUIRED for minimally competent practice – The REQUIRED option point value in the section equals MPL Presented at the 2004 CLEAR Annual Conference September 30 – October 2 Kansas City, Missouri Minimum Passing Level • DM MPL – The sum of all DM section MPLs • IG MPL – The sum of all IG section MPLS • Overall Simulation Problem MPL – Candidates must achieve MPL in both Information Gathering and Decision Making Presented at the 2004 CLEAR Annual Conference September 30 – October 2 Kansas City, Missouri Simulation Exam Development • 8 to 10 simulation problems per examination • Each problem assesses different situation typically encountered on the job Presented at the 2004 CLEAR Annual Conference September 30 – October 2 Kansas City, Missouri Let’s Attempt A Computerized Simulation Problem!!! Presented at the 2004 CLEAR Annual Conference September 30 – October 2 Kansas City, Missouri • Karen S. Flint, Director, Internal Development & Systems Integration Applied Measurement Professionals, Inc. 8310 Nieman Road Lenexa, KS 66214 913.541.0400 (Fax – 913.541.0156) [email protected] www.goAMP.com Presented at the 2004 CLEAR Annual Conference September 30 – October 2 Kansas City, Missouri “Practical Testing” Dr. Frank Hideg, DC Dr. Mark Christensen, PhD Dr. Paul Townsend, DC Presented at the 2004 CLEAR Annual Conference September 30 – October 2 Kansas City, Missouri NBCE History • The National Board of Chiropractic Examiners was founded in 1963 • The first NBCE exams were administered in 1965 • Prior to 1965 chiropractors were required to take chiropractic state boards and medical state basic science boards for licensure Presented at the 2004 CLEAR Annual Conference September 30 – October 2 Kansas City, Missouri NBCE Battery of Pre-licensure Examinations • • • • Part I – Basic Sciences Examinations Part II – Clinical Sciences Examinations Part III – Written Clinical Competency Part IV – Practical Examination for Licensure Presented at the 2004 CLEAR Annual Conference September 30 – October 2 Kansas City, Missouri Hierarchy of Clinical Skills DO PRACTICE SHOW HOW PART IV KNOW HOW PART III KNOWLEDGE PARTS I & II Presented at the 2004 CLEAR Annual Conference September 30 – October 2 Kansas City, Missouri NBCE Practical Examination Content Areas • Diagnostic Imaging • Chiropractic Technique • Chiropractic Case Management Presented at the 2004 CLEAR Annual Conference September 30 – October 2 Kansas City, Missouri Content Weighing TEC 17% DIM 16% DIM CAM TEC CAM 67% Presented at the 2004 CLEAR Annual Conference September 30 – October 2 Kansas City, Missouri Diagnostic Imaging • 10 Four-minute Stations • Candidate identifies radiological signs on plain film x-rays • Candidate determines most likely diagnoses • Candidate makes most appropriate initial case management decisions Presented at the 2004 CLEAR Annual Conference September 30 – October 2 Kansas City, Missouri Chiropractic Technique • 5 five-minute stations • Candidate demonstrates two adjusting techniques per station • Cervical spine • Thoracic spine • Lumbar spine • Sacroiliac articulations • Extremity articulations Presented at the 2004 CLEAR Annual Conference September 30 – October 2 Kansas City, Missouri Chiropractic Case Management • 10 five-minute patient encounter stations • 10 linked post-encounter probe (PEP) stations • Candidate performs focused case histories • Candidate performs focused physical examinations • Candidate evaluates patient clinical database • Candidate makes differential diagnoses • Candidate makes initial case management decisions Presented at the 2004 CLEAR Annual Conference September 30 – October 2 Kansas City, Missouri Key Features of NBCE Practical Examination • Use of standardized patients • Use of OSCE format and protocols Presented at the 2004 CLEAR Annual Conference September 30 – October 2 Kansas City, Missouri Case History Stations • Successful candidates use organized approach while obtaining case history information • Successful candidates communicate effectively with patients • Successful candidates respect patient dignity • Successful candidates elicit adequate historical information Presented at the 2004 CLEAR Annual Conference September 30 – October 2 Kansas City, Missouri Perform a Focused Case History Presented at the 2004 CLEAR Annual Conference September 30 – October 2 Kansas City, Missouri Post-Encounter Probe Station Presented at the 2004 CLEAR Annual Conference September 30 – October 2 Kansas City, Missouri Part IV Candidate Numbers 4500 4000 3500 3000 2500 2000 1500 1000 500 0 1996 1997 1998 1999 2000 2001 2002 2003 Presented at the 2004 CLEAR Annual Conference September 30 – October 2 Kansas City, Missouri Part IV State Acceptance 42 43 45 46 47 38 36 27 7 Colum n 1 1996 1997 1998 1999 2000 2001 2002 2003 2004 7 27 36 38 42 43 45 46 47 Presented at the 2004 CLEAR Annual Conference September 30 – October 2 Kansas City, Missouri All shaded areas are those states that ACCEPT or REQUIRE successful completion of the NBCE Part IV at the NBCE recommended passing level of 375 as one portion of that state’s licensure evaluation process. Those states are: Alabama Alaska Arizona Arkansas California Colorado Connecticut Delaware Georgia Hawaii Idaho Indiana Iowa Kansas Kentucky Louisiana Maine Maryland Massachusetts Minnesota Mississippi Missouri Montana Nebraska Nevada New Hampshire New Jersey New Mexico New York North Dakota Ohio Oklahoma Oregon Pennsylvania Rhode Island South Carolina South Dakota Tennessee Texas Utah Vermont Virginia Washington Wisconsin Wyoming * Those states requiring a passing score other than the NBCE recommended score are not listed. Presented at the 2004 CLEAR Annual Conference September 30 – October 2 Kansas City, Missouri Candidate Qualifications • Candidates must pass all basic science and clinical science examinations before applying • Candidates must be within 6 months of graduation from an accredited chiropractic college • $1,075 examination fee Presented at the 2004 CLEAR Annual Conference September 30 – October 2 Kansas City, Missouri Presented at the 2004 CLEAR Annual Conference September 30 – October 2 Kansas City, Missouri Contact Information • National Board of Chiropractic Examiners 901 54th Avenue Greeley, CO 80634 970-356-9100, 970-356-1095 [email protected] www.nbce.org Presented at the 2004 CLEAR Annual Conference September 30 – October 2 Kansas City, Missouri Station Development Alison Cooper Manager of Examination Operations Canadian Alliance of Physiotherapy Regulators Presented at the 2004 CLEAR Annual Conference September 30 – October 2 Kansas City, Missouri First Principles • If it’s worth testing, it’s worth testing well – it is possible to test anything badly – this is more expensive • Some things are not worth testing – trivia – infrequently used skills Presented at the 2004 CLEAR Annual Conference September 30 – October 2 Kansas City, Missouri Overview • • • • Write Review Dry run Approve Presented at the 2004 CLEAR Annual Conference September 30 – October 2 Kansas City, Missouri Write • • • • • • • Focus of station SP portrayal - general Checklist & scoring Instructions to candidate Details of SP instructions Review everything References Presented at the 2004 CLEAR Annual Conference September 30 – October 2 Kansas City, Missouri Focus of Station • Each station must have a clear focus – establish the focus in one sentence – take time to get this right – you can’t write a good station without a clear focus • Example: Perform passive range of motion of the arm for a client who has had a stroke. Presented at the 2004 CLEAR Annual Conference September 30 – October 2 Kansas City, Missouri SP Portrayal - General • Consider SP movement, behaviour – a picture in your head – use real situations to guide you • Not detailed yet • Example: Client is 55 years old, is disoriented, and has no movement in the left arm or leg. Presented at the 2004 CLEAR Annual Conference September 30 – October 2 Kansas City, Missouri Checklist & Scoring • • • • • What is important to capture Consider the level of the candidates Group items logically Assign scores to items Scoring scales Presented at the 2004 CLEAR Annual Conference September 30 – October 2 Kansas City, Missouri Checklist Example • • • • • • • • • • Explains purpose of interaction Corrects client’s position Performs passive ROM of scapula Performs passive ROM of shoulder Performs passive ROM of elbow Performs passive ROM of wrist Performs passive ROM of hand & fingers Performs passive ROM of thumb Uses proper body mechanics Uses proper handling 1 2 1 1 1 1 1 1 Presented at the 2004 CLEAR Annual Conference September 30 – October 2 Kansas City, Missouri 3 3 Instructions to Candidate • Information the candidate needs – age and sex of client – pertinent information and assumptions • The task for the candidate – exactly what they are to do and not do Presented at the 2004 CLEAR Annual Conference September 30 – October 2 Kansas City, Missouri Example Eric Martin 55 years old This client had a right middle cerebral artery haemorrhage resulting in a left sided hemiplegia two (2) weeks ago. The client presents with confusion and left sided flaccidity. His cardiovascular status is stable. • Perform passive range of motion on the client’s left upper extremity. • Perform only one (1) repetition of each movement. • Assume that you have the client’s consent. Presented at the 2004 CLEAR Annual Conference September 30 – October 2 Kansas City, Missouri Details of SP Instructions • History, onset, changes • Initial position, movements, demeanor, must say/ask – anticipate strong AND weak candidates • Cover the checklist and candidate instructions • SP prompts Presented at the 2004 CLEAR Annual Conference September 30 – October 2 Kansas City, Missouri SP Instructions... • Use plain language • Include – what to wear/not wear – features of the SP (height, scars) • Diagrams are often helpful Presented at the 2004 CLEAR Annual Conference September 30 – October 2 Kansas City, Missouri Example • • • • • • • • • • • Presenting complaint Initial position, general mobility, affect Comments you must make Medical, social history Medications Activities and areas affected Sensation Pain Muscle quality Responses to candidate Emotions Presented at the 2004 CLEAR Annual Conference September 30 – October 2 Kansas City, Missouri Check Everything • Go back and check – does it make sense? – is there still a clear focus? – is anything missing? • Edit/revise as needed – add notes to examiner for clarification • Check for plain language Presented at the 2004 CLEAR Annual Conference September 30 – October 2 Kansas City, Missouri References • Use references you expect candidates to know • Umphred, 2nd edition, page 681 Presented at the 2004 CLEAR Annual Conference September 30 – October 2 Kansas City, Missouri Next Steps • Review by others • Dry run • Approve for use Presented at the 2004 CLEAR Annual Conference September 30 – October 2 Kansas City, Missouri Thank you Canadian Alliance of Physiotherapy Regulators 1243 Islington Ave., Suite 501 Toronto, ON, Canada M8X 1Y9 (W)416-234-8800, (F)416-234-8820 [email protected] www.alliancept.org Presented at the 2004 CLEAR Annual Conference September 30 – October 2 Kansas City, Missouri “OSCE Research: The Key to a Successful Implementation ” Lila J Quero Muñoz, PhD Consultant Presented at the 2004 CLEAR Annual Conference September 30 – October 2 Kansas City, Missouri Prior to the OSCE: CPBC and PEBC • Need for assessing communication, counseling, and interpersonal skills to provide pharmaceutical care to patients • PEBC MC examination was not assessing the full scope of pharmacy practice as profiled by NAPRA (National Association Pharmacy Regulatory Authorities of Canada) Presented at the 2004 CLEAR Annual Conference September 30 – October 2 Kansas City, Missouri Generalizability: Data Analyses • Psychometrically, OSCEs, are complex phenomena, producing scores with potential errors from multiple sources, including: – – – – – Examiners (pharmacists and non-pharmacists) Cases (context, complexity, # of stations) Scoring methods (global vs. checklists) Standard setting Differential grading practices Presented at the 2004 CLEAR Annual Conference September 30 – October 2 Kansas City, Missouri Research Question # 1 • How many examiners are required to obtain consistent and dependable candidates’ scores? Presented at the 2004 CLEAR Annual Conference September 30 – October 2 Kansas City, Missouri Results #1-1998 • 1 examiner per case yielded similar consistency as 2 (G=.82, .81, D=.81, .79) indicating that examiners agreed highly on their scores • Examiners contributed little to the scoring errors of candidates’ performance Presented at the 2004 CLEAR Annual Conference September 30 – October 2 Kansas City, Missouri 1 vs. 2 Global -1999 Global Grading Roving & Assessor G and D Reliability as a Function of Stations & Raters Mean of 3 Sets of Stations G & D Reliability Estimates G Rel 1Rater G Rel 2 Raters 1 D Rel 1 Rater 0.9 D Rel 2 Raters 0.8 0.7 0.6 0.5 0.4 0.3 3 8 10 13 15 17 20 # of Stations Presented at the 2004 CLEAR Annual Conference September 30 – October 2 Kansas City, Missouri Research Question # 2 • How many cases are required to maintain consistency, validity and generalizability of scores? – Adequate and representative sampling of professional practice are necessary to capture a candidate’s abilities. – Multiple observations of abilities yield more consistent and content valid inferences. – Logistical constraints restrict the number of cases that are timely and economically feasible to administer within one OSCE examination. Presented at the 2004 CLEAR Annual Conference September 30 – October 2 Kansas City, Missouri Results # 2-1998 • 15 cases reduced the candidate’s score error due to sampling variability of the cases dramatically from 5 or 10 cases and improved the consistency of scores from G=.60 to .81 • 15 cases reduced the cases and raters interaction variance as an indication that raters agreed on their scores across cases Presented at the 2004 CLEAR Annual Conference September 30 – October 2 Kansas City, Missouri Results # 2-1998 • Candidates’ scores varied mostly due to their differential performance across cases. • Sampling of the cases might affect the candidates’ performance on an OSCE. • We suggest, however, that differential performance across cases might be due to candidate’s differential levels of skills across the pharmacy competencies assessed Presented at the 2004 CLEAR Annual Conference September 30 – October 2 Kansas City, Missouri Profile of Sources of Errors in %-1998 45 True score variance 40 35 30 25 20 15 10 5 0 Rater variance Case variance Interaction can. & rater Interaction can. & case Interaction rater & case Presented at the 2004 CLEAR Annual Conference September 30 – October 2 Kansas City, Missouri Research Question # 3 • How do different scoring methods such as checklists or global grading affect candidates’ scores? Presented at the 2004 CLEAR Annual Conference September 30 – October 2 Kansas City, Missouri Results # 3-1998 • Low correlations between checklist and global scores suggest both methods might not be interchangeable • If used in isolation they would yield different end results, particularly for borderline candidates • Global grading yields higher mean scores than checklist grading (p values.81 and .59) Presented at the 2004 CLEAR Annual Conference September 30 – October 2 Kansas City, Missouri Global vs. Checklist-1999 C o mp a r i n g H o l i s t i c ( G l o b a l ) & C h e c k l i s t G r a d i n g G a n d D G &D Estimates R e l i a b i l i t y ( A s s e s s o r & St a n d a r d i z e d P a t i e n t ) / O n e R a t e r 0.9 0.7 0.5 0.3 8 10 12 14 16 18 20 Stations Presented at the 2004 CLEAR Annual Conference September 30 – October 2 Kansas City, Missouri G1 Check Nest Rel. D1 Check Nest Rel. G1 Global Nest Rel. D1 Global Nest Rel. Research Question # 4 • What is the validity and defensibility of standard-setting procedures and pass/fail decisions Presented at the 2004 CLEAR Annual Conference September 30 – October 2 Kansas City, Missouri Results # 4-1998 • SME’s agreed highly on the minimum standard necessary for safe pharmacy practice for the borderline qualified pharmacists • On different occasions, SME’s had similar standards for entry-to-practice for the core cases • Standards varied little between 26 & 20 cases and were consistent enough with 15 cases (G=.74, .74, .71) Presented at the 2004 CLEAR Annual Conference September 30 – October 2 Kansas City, Missouri Results # 4-2003 Standard Setters (SS) Variation Within Station and Mean 4.00 ss1 3.00 ss2 2.50 ss3 2.00 ss4 1.50 ss5 1.00 ss6 ss7 0.50 ss8 0.00 ss9 J0 10 J0 72 G0 08 J0 58 Ri 30 B Ro 09 J0 73 J0 68 J0 44 J0 42 J0 37 Q0 12 J0 65 P0 06 W 00 1 M ea n SS Mean Scores 3.50 Stations Presented at the 2004 CLEAR Annual Conference September 30 – October 2 Kansas City, Missouri ss10 MEAN Research Question # 5 • Are there differential grading practices among Canadian Provinces? • Are candidates’ pass/fail decisions affected by provincial differences on scoring practices? Presented at the 2004 CLEAR Annual Conference September 30 – October 2 Kansas City, Missouri Results # 5-Videos 2003 • Variability in scores between sites are due mostly to true score variance • Differences between exam sites are in magnitude of scores but not in pass/fail status • Differences between assessors are mostly of magnitude of scores but not in pass/fail status • Pass/Fails decisions did not vary between sites and assessors • There is more variance between assessors than between exam sites Presented at the 2004 CLEAR Annual Conference September 30 – October 2 Kansas City, Missouri Results # 5-2003 Stations J03 & J29 Spring 03 Assessors' Mean Scores 4 3.5 3 J03 2.5 J29 2 1.5 1 101 103 401 402 701 OSCE Sites and Assessors Presented at the 2004 CLEAR Annual Conference September 30 – October 2 Kansas City, Missouri Results # 5-2003 Global Grading Spring Exam Sites 10, 40 Global Grading Spring Exam Sites 10, 40 J03 J29 Facets df var SE P 18 0.150 0.077 L 1 0.058 R:L 2 PL PR:L Total Facets df var SE 25% P 14 0.267 0.104 69% 0.069 10% L 1 0.000 0.015 0% 0.029 0.035 5% R:L 2 0.038 0.030 10% 18 0.000 0.066 0% PL 14 0.026 0.021 7% 36 0.368 0.085 61% PR:L 28 0.057 0.015 15% 100% Total 0.606 %var 0.388 Presented at the 2004 CLEAR Annual Conference September 30 – October 2 Kansas City, Missouri %var 100% Results # 5-2003 Assessors' Mean Scores 124 Videos (6 Stations Spring 03 Exam) Scored by 35 Raters from 9 Exam Sites (Passing Score 2.786) 4 3.8 3.6 3.4 3.2 3 2.8 2.6 2.4 2.2 2 1.8 1.6 1.4 1.2 1 MJ33 MRa02 MJ38 MJ45 MJ03 Stations Presented at the 2004 CLEAR Annual Conference September 30 – October 2 Kansas City, Missouri MJ29 Conclusions 1998-2004 • Development of cases should follow templates, guidelines and a detailed blueprint • Selection of cases must follow a detailed blueprint to mirror OSCE forms between exam administrations to control for differences in cases such as complexity and content Presented at the 2004 CLEAR Annual Conference September 30 – October 2 Kansas City, Missouri Conclusions 1998-2004 • Multiple sources of errors in OSCEs forces us to do more extensive and nontraditional research than for MC exams • OSCEs require continuous vigilance to assess the impacts of the many sources of errors • OSCE research must be planned and implemented beyond exam administrations Presented at the 2004 CLEAR Annual Conference September 30 – October 2 Kansas City, Missouri Conclusions 1998-2004 • OSCE infrastructure must support both design research and exam administration research • Successful implementation and continuous improvements of OSCE go hand and hand with research • More collaborative efforts among OSCE users are needed to built on each other’s success and avoid pitfalls Presented at the 2004 CLEAR Annual Conference September 30 – October 2 Kansas City, Missouri Conclusions 1998-2004 • Although OSCE research is costly it is a deterrent to litigation and wasted exam administration resources • Similar conclusions may apply to other performance assessments Presented at the 2004 CLEAR Annual Conference September 30 – October 2 Kansas City, Missouri • Carol O’Byrne, BSP, OSCE Manager • John Pugsley, PharmD, Registrar, PEBC [email protected] 416-979-2431, 1-416-260-5013 Fax • Lila J. Quero-Muñoz, PhD, Consultant 787-431-9288, 1-888-663-6796 Fax [email protected] Presented at the 2004 CLEAR Annual Conference September 30 – October 2 Kansas City, Missouri