Read the article below and provide EXPLANATIONS OF YOUR ANSWERS by doing a screenshot of the answer from the article proving your answer is correct.

Question 1 16.67 pts Aims of this study include: Determining the likelihood of patients returning for follow up visits within two years of discharge from rehabilitation. Determining any risk factors associated with increasing a patient's risk of returning for a follow up visit by twelve months after discharge from rehab. Determining risk factors associated with societal participation, activity, function, health disorder and personal behaviors and their potential effect on returning for a follow up visit by twelve months after discharge from rehab. Determining any risk factors associated with increasing a patient's risk of returning for a follow up visit within two years of discharge from rehab. Question 2 16.67 pts All of the following persons were excluded in the study except: Persons with psychiatric problems. Persons classied as AD on the ASIA scale and were wheelchair bound. ' All of the persons described in these answers were excluded from the study. ' Those who had progressive disease. Question 3 16.67 pts What was the study outcome? ' A return for evaluation twelve months post discharge ' A signicant change in two or more clinical measurements at the follow up exam. ' The impact of a follow up exam twelve months post discharge. ' A return for evaluation two years post discharge. Question 4 16.67 pts After identifying the attrition bias, the authors determined the persons who dropped out differed to the participants who remained according to all of the following, except: " Levels of %FEV1 ' Age ' Level of traumatic injury. Gender Question 5 16.67 pts Exposure was measured by only clinical exam. True False Question 6 16.67 pts Since attrition is the loss of participants in a study, attrition bias will always underestimate the effect measure. True ' ' False J Rehabil Med 2009; 41: 382-389 ORIGINAL REPORT INVESTIGATION OF BIAS DUE TO LOSS OF PARTICIPANTS IN A DUTCH MULTICENTRE PROSPECTIVE SPINAL CORD INJURY COHORT STUDY Sonja de Groot, PhD12, Janneke A. Haisma, MD, PhD3, Marcel W. M. Post, PhD4, Floris W. A. van Asbeck, MD, PhD4 and Lucas H. V. van der Woude, PhD21 From the 'Rehabilitation Center Amsterdam, Research Institute MOVE, Faculty of Human Movement Sciences, VU University Amsterdam, Amsterdam, Department of Rehabilitation Medicine, Erasmus Medical Center Rotterdam, Rotterdam and 4Rehabilitation Center De Hoogstraat and Rudolf Magnus Institute for Neuroscience, University Medical Center Utrecht, Utrecht, The Netherlands Objective: To determine bias due to loss of participants (at- activity and societal participation, and the relationships be- trition bias) in a prospective cohort study. tween these levels. Each domain of the International Classifi- Design: A multi-centre prospective cohort study. cation of Functioning, Disability and Health (ICF) model (2) Subjects: A total of 225 individuals with a spinal cord injury is represented by several (sets of) outcome measures (Fig. 1). from 8 Dutch rehabilitation centres. To study restoration of mobility, a wide variety of physical Methods: Participants were considered non-participants tests and questionnaires were administered to recently injured when no information was collected at the measurement one patients with SCI at the start and discharge of inpatient reha- year after discharge from inpatient rehabilitation. Using bilitation and one year after discharge bivariate tests participants and non-participants were com- A problem in all longitudinal studies is loss of participants pared regarding personal, lesion, function and functional characteristics determined at the beginning of inpatient re- to follow-up, i.e. the response rate decreases between the first habilitation and at discharge. A logistic regression was per- measurement and the follow-up assessments (3, 4). In our pro- formed to determine which characteristics predict participa- spective cohort project 225 persons were included at the start of tion at one year after discharge. rehabilitation; however, at the last measurement (one year after Results: Of the participants at the start of the study, 31% discharge), only 156 persons performed the tests. This might lead (n=69) did not perform the tests one year after discharge to attrition bias, when the participants who participated in the from inpatient rehabilitation. Variables associated with last measurement are systematically different from those who did study participation one year after discharge were: higher not (4). This potential attrition bias has not yet been evaluated level of education, higher well-being score at the start of re- in our prospective cohort study, but might be very important habilitation, and a shorter length of stay in hospital and re- for 2 reasons. Firstly, since attrition will most likely occur in habilitation centre at discharge of inpatient rehabilitation. future longitudinal rehabilitation studies, it is important to gain Conclusion: Selective attrition in the longitudinal study more insight into the factors that determine whether participants might have led to an over-estimation of some of the results attend on recurring test occasions. Knowing which factors play of the measurement one year after discharge from inpatient a role might enable researchers to minimize loss of follow-up. rehabilitation. Secondly, it is important to investigate differences between Key words: spinal cord injuries, rehabilitation, prospective participants and non-participants at the measurement one year studies, follow-up studies, patient dropouts. after rehabilitation, because it might influence the results and J Rehabil Med 2009; 41: 382-389 generalizability of the conclusions of this cohort study. Since the outcome variables of the Dutch cohort study con- Correspondence address: Sonja de Groot, DNO, Revalidatie- cerned all domains of the ICF model, possible differences be- centrum Amsterdam, Overtoom 283, NL-1054 HW Amsterdam, The Netherlands. E-mail: s.de.groot@fbw.vu.nl tween participants and non-participants one year after discharge from inpatient rehabilitation also need to be investigated for all Submitted April 24, 2008; accepted November 26, 2008 ICF domains. For example, participants with a higher level of education might return to work (domain societal participation) INTRODUCTION earlier (5) and, therefore, might not be able or motivated to return to the rehabilitation centre for a research project. This could lead The Dutch prospective cohort project "Restoration of mobility to an under-representation of persons returning to work at the in spinal cord injury rehabilitation" is a multi-centre longitu- measurement one year after discharge. Another possibility is dinal cohort study conducted in 8 rehabilitation centres with that persons with a more severe SCI (domain health disorder), specialized spinal cord injury (SCI) units (1). The theoretical leading to more physical strain during daily activities, might not relevance of this cohort study is to improve our understanding be able to travel to the rehabilitation centre to perform all the of recovery of mobility at the levels of function and structure, tests one year after discharge. Based on the population actually J Rehabil Med 41 2009 The Authors. doi: 10.2340/16501977-0346 Journal Compilation 2009 Foundation of Rehabilitation Information. ISSN 1650-1977Attrition bias in a longitudinal SCI study 383 384 S. de Groot et al. Health status in Spinal Cord Injury Table I. Reasons for not performing the measurement one year after discharge from inpatient rehabilitation classified in 5 categories: (1) deceased, (2) refusal/did not attend, (3) could not be reached, (4) organizational reasons, (5) excluded during the study Disorders, diseases: spinal cord injury, secondary complications, co-morbidity Non-participants Non-participants Participants t discharge (n) one year after discharge (n) one year after discharge (n) Total Participants one year after discharge 156 156 Reasons for non-participation one year after discharge Drop-outs Functions and anatomical Activities: Hand- & Participation: . Decease properties: Cardiovascular arm functionality, Work, school, 2. Refusal & respiratory function, Did not attend at appointment UNNG O owoou musculoskeletal function, Basic & complex sports, family, motor & sensory functions (wheelchair) skills, friends Transferred/moved ADL independence Could not be reached Not invited 4. Contract research assistant ended Can walk again NN OO - Fig. 1. (Sets of) outcome measures in the Dutch Psychiatric problems after the start of rehabilitation cohort study for the different domains and contexts Total of the International Classification of Functioning, 2 of these participants did not participate at discharge but came back one year after discharge. Environmental factors Personal factors Disability and Health (ICF) model as were among Rehabilitation treatment, Age, gender, education others evaluated in the Dutch cohort study. ADL: strategy, practice, activities of daily living. To determine the strength of the upper extremities, the shoulder ab- Statistics ductors, internal and external rotators, elbow flexors and extensors, and wrist extensors in both arms were tested Scores at the start and discharge of inpatient rehabilitation of participants Outcome measure and determinants (MMT). This test was performed in standardized positions, in which and non-participants at the mea ment one year after inpatient rehabili- measured, the results of the domain function and activities may participants performed a movement either with gravity eliminated, tation were compared using independent samples t-tests (parametric data) then be too positive at one year after discharge. The outcome of interest was the actual study participation at the measure- against gravity, or against resist rated on a scale and Mann-Whitney U tests or y" tests (non-parametric data). Differences Therefore, the objectives of this study were to investigate: (i) ment one year after discharge from inpatient rehabilitation. Discrimina ranging from 0 to 5 (9). Summing the scores of the 12 muscle groups in participation b th a Kruskal-Wallis possible differences in characteristics related to personal (age, tive analyses and a pr ostic model were used o examine whether gave an MMT sum score (maximum 60). test (p0.05), in order to registered each complication (pressure sores, respiratory infection, very unsatisfying, unsatisfying, rather unsatisfying, rather satisfying, create the final multivariate model. d the satisfying and very satisfying. This score was dichotomized (unsatis- The regression coefficients for the determinants were converted to purpose of the study and the testing methods. Since restoration of mo- urinary tract infection) on a standardized list (7). At the start of active bility in most people with SCI means using a wheelchair, only subjects rehabilitation, com occurred since fied: 1-4 vs satisfied: 5-6) according to Fugl-Meyer et al. (14). odds ratios (ORs). An OR of 1 indicated that there was no association who were, at least in part, wheelchair-dependent re included. admission to rehabilitation were registered. At discharge from inpatient All tests and protocols were approved by the medical ethics commit- rehabilitation, complications then present or that had occurred over the ee. After they had been given information about the testing procedure, last 3 months were registered. The occurrence of a complication was Start Participants n = 225 all participants completed an informed consent form. registered as follows: 0=no complication; 1 = one or more complica- tions are or were present (7). Design Domain function. To assess respiratory function, flow-volume curves Data for the current study were collected at the start of active rehabili- Discharge Participants Non-Participants were prepared with an Oxycon Delta (Jaeger, Hoechberg, Germany). n = 198 n = 27 tation when the participant could sit in a wheelchair for at least 3 h, at The forced expiratory flow/sec (FEV1) (in 1) for each participant was discharge from inpatient rehabilitation, and one year after discharge. expressed as a percentage of the expected score for that participant Fig. 2. Participants and loss to follow- Data were collected by trained research assistants with a paramedic in comparison with an age-, sex-, and height-matched able-bodied 1 year after Participants Non-Participants Participants Non-Participants up in the Dutch prospective cohort background using standard procedures and equipment. population (8). discharge n = 154 n = 44 n = 2 n = 25 study (1). J Rehabil Med 41 J Rehabil Med 41Attrition bias in a longitudinal SCI study 385 386 S. de Groot et al. Table II. Descriptive characteristics at the start and discharge from inpatient rehabilitation of the participants and non-participants one year after Table III. Results of the final multi-level logistic random coefficient Table IV. Results of the final multi-level logistic regression model with discharge from inpatient rehabilitation (n = 225) model with the outcome variable participants (= 1) vs non-participants the outcome variable participants (=1) vs non-participants (=0) and Non-participants, n=69 Participants, n= 156 (=0) and the association with significant independent variables the association with significant independent variables measured at measured at the start of active rehabilitation discharge from inpatient rehabilitation Variable % or mean (SD) % or mean (SD) Beta (SE) Odds 95% CI P Beta (SE) Odds 95% CI P Personal characteristics Age, years 67 42.2 (14.2) 155 39.8 (14.0) 0.25 Constan 0.249 (0.254) Constant 1.350 (0.410) Gender Education dummy 0.578 (0.392) 1.78 0.83-3.84 0.14 men 12 0.18 Education dummy 1 0.586 (0.368) 1.79 0.87-3.70 0.1 Education dummy 2 1.010 (0.445) 2.75 1.15-6.57 0.02 women 10 13 28 Education dummy 2 0.800 (0.391) 2.23 1.03 4.79 0.04 Time since injury a Disorders, diseases discharge (months) -0.070 (0.029) 0.93 0.88 0.99 0.02 Cause of injury Well-being at start 0.785 (0.394) 2.30 1.01-4.75 0.04 Education dummy 1: low (1) vs mid (0) level of education; Education traumatic 47 69 17 76 0.33 Education dummy 1: low (1) vs mid (0) level of education; Education dummy 2: mid (0) vs high (1) level of education; SE: standard error non-traumatic dummy 2: mid (0) vs high (1) level of education. Well-being: 95% CI: 95% confidence interval. Lesion level at start unsatisfied (0) and satisfied (1); SE: standard error; 95% CI: 95% tetraplegia 30 44 36 0.23 confidence interval. paraplegic 100 Lesion level at discharge education. Those persons who were satisfied with their life at tetraplegia 25 42 47 35 0.42 the start of active rehabilitation were 2.3 times more likely to between the independent variable and study participation one year paraplegic after discharge, a significant OR > 1 indicated an increased chance of attend one year after discharge from inpatient rehabilitation, Completeness at start study participation one year after discharge from inpatient rehabilita while those who had a one month longer length of stay for complete 38 110 29 0.21 tion, whereas a significant OR