Article Appraisal WorksheetArticle #___2__Article Appraisal Worksheet Page| 4Section 2: Critical Analysis Are the results of this individual study valid?Question to ConsiderResponse(use supporting information from the article in your response)What type of study design was utilized for this study? Is that the optimal type of design for the purposes of the study?-The study design used in this study was aninception cohort study. -Yes,I believe this approach was an idealapproach for this study becauseit allowed the researchers to gatherindividualsat a time early in the development of a specific clinical disorderwho are followed thereafter. Was there anyrandomizationprocess? If so, how was randomization handled?If no randomization, how could that impact the results reported? Is that appropriate for this type of study?-Yes, randomization was utilized in this research study. It was determined that randomizationwas computed with a random –number generator performed by an independent research (ACV) who was not responsible for determining eligibility.-Yes, I believe this was appropriate for this type of study because it allowed the study to be completely randomized for which group they were placed inwhich will aid in minimizing bias.Was follow-up of patients/length of study sufficiently long and complete?Consider time needed for the conditions of interest to develop.-The patients were followed-up at week 4,12, and 52 which is long and complete. The long follow up allows for an sufficientamount of time to follow the patients, especially those who hadchronic conditions.Were all patients analyzed in the groups to which they were originally allocated?Intention to treat? If no, what did the authors state about how drop-outs were handled?-Yes,the patients were analyzed in the group they were assigned which was determined at week 4 and whether they had “recovered” or “not recovered”.-Several patients were initially omittedfrom the study for not meeting certainstandards, which include:refusing to participate, and/ or already receiving treatment prior to the study. -Furthermore, 6 patients who were in the study at weeks 4 and 12 withdrew prior to week 52. The reasons for withdrawal included: loss of motivation, recovered and not willing to complete the follow-up measurements,not having enough time to participate in the study,or some patients
Article Appraisal WorksheetArticle #___2__Article Appraisal Worksheet Page| 5were not satisfied with the treatment they received. However, there is no intention to treat because this study is a prognosis.Was the prognostic finding evaluatedin an appropriate spectrum of patients?Was there any selection bias?-Some of the prognostic findings include: high initial pain intensity is a prognostic factor for a poor recovery, and the pain and activities measured by a VAS have a high prognostic value.These prognostic tools are appropriate for the study because they cover a wide spectrum of patients. -Yes, there was selectionbias in the study. Some factors that could lead to this selection bias were: researchers had a difficult time recruiting whiplash patients and they were hoping to recruit more patients but were unable to. Lastly, the design of their study could lead to selection bias.Additionally, it is not said if the samplesize was suitablein this study. Were groups treated equally throughout the study?Was there any “work-up” bias?-The groups were not treated equally throughout the study, but patients were treated equally within each groupthey were assigned. -One group was placed in the care of a physical therapist-whose treatment consisted of patient education, advice, graded activity, and exercise therapy. The other group was placed in the care of a general practitioner, whose treatment consisted of patient education and advicefor recovery. The maximumduration of the treatment in the study was 9 months but there was no limit on the amount of times the patients had to receive treatment.-Lastly,there was work-up bias in the study. The length of the sessions and hands-on time were different between groups. Thetreatment from the physical therapist were 30 minutes in duration and included hands on time. Whereas, the general practitioner’streatmentwasonly 10minutes per session and did not include any hands-ontime. Were the groups similar at the start of the study? If different, what could the impact of those differences be on the results reported?-Yes, the groups were similar at the start of the study. -Originally, in the first phase of the study patients were recruited by general practitioners. Next, physicians examined each patient to determine if they met the specific criteria of the study.
Article Appraisal WorksheetArticle #___2__Article Appraisal Worksheet Page| 6-The following prognostic factors were taken into consideration: socio-demographic characteristics (age, gender, education, maritalstatus, health insurance, and employment status), crash-related characteristics (direction of impact, location in vehicle, seatbelt use, unprepared for collision, head restraint and correctly positioned, traffic situation), pre-existing (comparable) health problems (previous neck pain, headache, participation problems, co-morbidity and pain medication use before injury), physical health characteristics 2 weeks following the collision, psychological symptoms measured with the SCL-90, and other factors(such as additional imaging or attained a lawyer).-One impact that could have played a factor in the results is if there weren’t such specific criteria the patients had to meet to be included in the study.Had more patients been included in the study,the prognostic factors would not produce the same outcomes,and/or the validity of the prognostic factors would be different. Was the diagnostic test/prognostic factor validated in a second, independent group of patients?-The author does not make a note if an additional study was conducted to determine if the prognostic factors were validated in a second, independent group of patients. -However, the author does mentionthat future research needs to be doneto see if the same prognostic factors are relevant for other populations.Are the valid results of this individual study important?Are the findings of the studyof significance?-Yes, at weeks 4, 12, and 52, neck pain intensity and work activitiespredicted a significant difference in rate of recovery. These two factors contributed the most to the total explained variance. -However, the impact of the interventions was only significant at 12 weeks, showing a strong negative association with poor recovery. -The VAS is both a valid and reliable measure to assess pain and level of function post-injury.
Article Appraisal WorksheetArticle #___2__Article Appraisal Worksheet Page| 7Also, the SCL-90 has a high face validity, test-retest correlation, and internal consistency. Did the prognostic factor demonstrate an ability to differentiate differences between those patients with and without the disorder?-Yes, the accuracy of the predictions of the prognostic factors was high,which means the prognostic factors were an accurate predictor of how well patients recovered from their injury. Some prognostic factors had a greater correlation with recoverythan others. For example, neck pain intensity and work disability were themost significantpredictors. On the other hand, a prognostic measure that wasn’t as significant was female gender at the end of the study.How was the prognostic finding analyzed? What statistical analysis approach was utilized?-The prognostic findings were analyzed with logistic regression models created with theuse of baseline variable as possible prognostic factors for faster recoveries. The results of the analysis at weeks 12 and 52 were altered for the treatment effects with a three-stage modeling procedure.-At week 52 the only factors of interest with P-values less than 0.05 were included and the resultsthatwere presented as unstandardized regression coefficients and odds ratios with 95% confidence intervals. -The area under the curve is understoodas the probability of correctly identifying a patient with poor recoveryfrom randomchosen patients, where 1.0 is a perfect discrimination between two health studies. -Also, the P-values in the study were two-tailed.Was the result of the analysis statistically significant? For which findings? For which groups?-After 4 weeks, 7 of the 18 factors which were related topoor recovery were included in the multiple logistic regression model.-Neck pain intensity (B = 0.041, P-value = 0.001), somatization (B = 0.15, P-value = 0.002), and sleep difficulties(B = 0.405, P-value = 0.003) showed higher initialscoresand lower initialscores for work-related activities (B = -0.034, P-value = 0.000) were associated with poor recovery. -After 12 weeks, 9 of the 19 factors related topoor recovery were included in the multiple logistic regression model. Female gender(B = 0.977, P-value = 0.017), being unprepared for collision(B = 0.882, P-value = 0.064), neck pain intensity (B
Article Appraisal WorksheetArticle #___2__Article Appraisal Worksheet Page| 8= 0.041, P-value = 0.001), somatization (B = 0.087, P-value = 0.008), and sleep difficulties(B = 0.041, P-value = 0.001), while a lower initialscore for work related activities (B = -0.022, P-value = 0.000) were associated with a poor recovery-After 52 weeks, 10 of the 19 factors relatedwith poor recovery, were included in the final multiple logistic regression model. These factors included poor recovery was predicted by female gender (B = 1.553, P-value = 0.001), and a low level of education (B = 1.390, P-value = 0.007), and higher baseline scores for neck pain intensity (B = 0.028, P-value = 0.006), and somatization (B = 0.19, P-value = 0.009), and lower baseline scores for work-related activities (B = -0.014, P-value = 0.009). The predictive power of these factors correctly classified 76.8% with an explained variance of 45.9%. The accuracy of the model was modest to high with an AUC of 0.86 (95% CI: 0.88; 0.97). How do you know that the results were statistically significant? -This study used B values, which is a regression coefficient that tells about the change in the value of the dependent variable and how it corresponds to the change in the independent variable. -The B-values representedthe difference as a patient progressedduring therecovery period. The variables in the study that characterizedmore intense pain/ disabilities at the start had a more positive B value, or a reduction of pain symptoms. For the patients whose symptoms got worse, as the study went on had negative B values.-The area under the curve gavean indication of the predictive power and accuracy of the models the percentage of the explained variance (Nagelkerke’s R2). -Because the data at weeks 4, 12, and 52 were high, it can be determinedthat patients will have
Article Appraisal WorksheetArticle #___2__Article Appraisal Worksheet Page| 9better outcomes by finding these prognostic factors who suffered acute whiplash injuries.-Finally, the research models were adjusted by the general practitioner and the physical therapist at weeks 12 and 52. The impact of the interventionswere only significant at week 12, which shows a strong negative association with poor recovery.-The odds ratio, which was 0.363, showsthat patients of the general practitioner or physical therapist led to better outcomes than the study was expected.-An OR of 0.363 means the outcome was 63.7% less likely to occur (Szumilas, M, 2010). This statistic was proven to be significant because P = 0.008.Were the results for thestudy“clinically significant?”-The regression analysis in the study showsthe prognostic factors with a low initial score for work activities and a high initial score for neck pain intensity at weeks 4, 12, and 52 associateswith a poor, or slow, recovery outcome for the patients inthe study.-Additionally, by week 12, interventions under the care of the general practitioner or physical therapist showed to have better outcomes for the patients.-Three factors:female gender, low level of education, and sleep difficulties, were statistically significant at the beginning and end ofthe study which can provide future guidance to ensure early intervention and better outcomes. Is this estimate of the of the diagnostic test/prognostic finding precise?(95% CI) Interpret the confidence intervals.-The confidence intervals worth notinginclude:neck pain intensity and work activities. -Neck pain intensity falls within the confidence interval and is above 1.00. So,it can be determinedthat there is more than a 95% chance of the patient having the intended outcome. The dataat week 4 had an odds ratio of1.034 with a CI of 1.006; 1.062, week 12 had an odds ratio 1.025 with a CI of1.002; 1.049, and 52 weeks with an odds ratio of 1.020 with a CI of 1.002;1.038.
Article Appraisal WorksheetArticle #___2__Article Appraisal Worksheet Page| 10-However, work activities odds ratios fell within the confidence interval,so it can be surmisedthat there is less of an association between the initialscores, leading to a poor recovery. Are there any “gold standard” measures (for comparison) used in this estimation?Did they have any comparison measure?-The authors of the study used the results they gathered and compared them to three systematic reviews. Two systematic reviews these authors compared their findings towere done by Schloten-Peeter et al. in 2003. Schloten-Peeter conducted research on different prognostic factors associated with whiplash injuries and a study done with whiplash patients and conservative treatment. A third study used for comparison was done by Cote et al in 2001 is titled: A systematic review of the prognosis of acute whiplashand a new conceptual framework to synthesize the literature.-The inconsistent results on the course of whiplash in the study populations are probably due to difference in study populations and definitions of outcome. Are these valid, important resultsapplicable to my question?Is the diagnostic test/prognostic factorfrom the study feasible to use/consider in the clinical setting?-Yes, the intervention from the study is feasible in the clinical setting. The study can be easily replicated especially because the VAS pain score and activity score can easily be administered to patients. -Also, the study provides a thorough description of the study design and interventions which can be altered for each individual patient, if need be.Does the prognostic finding aid in the clinical decision-making processes utilized clinically? -Yes, the prognostic factors aid in the clinical decision-makingprocesses utilized clinically. -The prognostic factors related to a poor recovery include: female gender, a low level of education, high initial neck pain, more severe disability, higher levels of somatization and sleep difficulties.These prognostic factors can guide the clinician how to structure their treatmentfor that
Article Appraisal WorksheetArticle #___2__Article Appraisal Worksheet Page| 11specific patientto ensure no patient has a poor recovery. Is the population being studied similar enough to the individual(s) I have seen in clinical settingsto try and use this test, or measure this factor?-Yes, the population included in the study were patients from ages 18 –55. The patients studied had grad 1 or 2 WAD resulting from a traffic accident including:symptoms of neck pain, headache, or dizziness within 48 hours after trauma.-This study is similar to patients I have seen in clinic thus far and will continue to see. The only difference would be that this study was conducted in the Netherlands.Are there any limitations/considerations apparent in this study?(Hint: Bias, Error)-Yes, there area couple limitations within the study. The author stated how bias may have been involved because of the difficulty in recruiting whiplash patients. Also, the study design that was utilized may have also lead to selection bias. -Another source of possible bias in the study is that the models were over fitted. The researches included many variables compared to the frequency of the outcome, which decreases the stability of the model.-The authors stated the application of their models in different populations and settings is necessary to determine if they can be generalized to the greater public because they may have produced lower values for accurate predictions among patients in other populations. -Future research will need to be conducted to confirm these findings andshould focus on thresholdpoints for high neck pain intensity and low levels of activities on the VAS. How will the results of this study help me in caring for my patients?-It was determined by Erik et al. that baseline prognostic factors for patients with symptoms of whiplash associated disorders grade 1 or 2 lead to poor patient outcomes with high initial neck pain and lower initial scores for work activities. The prognostic factors include: female gender, a low level of education, high initial neck pain, more severedisability, and higher levels of somatization and sleep difficulties. The results of this study will help me in caring for my future patients because I will be able to utilize the visual analog scale, as itwill not take long for the patient to fill out and will provide me with quality information to structure my treatment for the patient to ensure an adequate recovery time.
Article Appraisal WorksheetArticle #___2__Article Appraisal Worksheet Page| 12
Article Appraisal WorksheetArticle #___2__Article Appraisal Worksheet Page| 13Additional CitationsSzumilas, Magdalena. "Explaining Odds Ratios." Journal of the Canadian Academy of Child and Adolescent Psychiatry. August 2010. Accessed April 22, 2018. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2938757/.
11 / 13