Do Cervical Spine Surgery Patients Recall Their Preoperative Status?: A Cohort Study of Recall Bias in Patient-reported Outcomes

Recall Bias in Client-Reported Outcomes in Canine Orthopaedic Patients Using Clinical Metrology Instruments

OBJECTIVE

 The aim of this study was to determine the accuracy of client recollection of their dogs' preconsultation status using clinical metrology instruments such as the Liverpool Osteoarthritis in Dogs (LOAD) and Canine Brief Pain Inventory (CBPI) questionnaires in dogs presenting to a referral orthopaedic clinic.

STUDY DESIGN

 This is a longitudinal prospective cohort study of client-owned dogs presenting for investigations of lameness (n = 217). LOAD and CBPI questionnaires were completed by the owners at the first consultation (T0). Owners were contacted at 2 (T1), 6 (T2), and 12 (T3) months and asked to recall their dogs' T0 status by completing another LOAD and CBPI questionnaire. The agreement between the T0 and recalled LOAD and CBPI scores was determined using the two-way mixed effects intraclass correlation coefficient (ICC). The Wilcoxon signed-rank test was used to determine the difference between scores.

RESULTS

 For the LOAD scores, there was moderate agreement between T0 and T1 (ICC: 0.64) and T0 and T2 (ICC: 0.53) scores and poor agreement between T0 and T3 (ICC: 0.496). For the CBPI Pain Severity Scores, there was poor agreement between T0 and all three subsequent time points (ICC < 0.5). For the CBPI Pain Interference Scores, there was moderate agreement between T0 and T1 (ICC: 0.57) and T2 (ICC: 0.56) scores and poor agreement between T0 and T3 (ICC: 0.43).

CONCLUSION

 The LOAD and CBPI questionnaires are subject to recall bias. Studies reporting retrospectively acquired CMI data should be interpreted with caution.

Do Patients Accurately Recall Their Preoperative Symptoms After Elective Orthopedic Procedures?

Patient-reported outcome measures are a frequent tool used to assess orthopedic surgical outcomes. However, recall bias is a potential limitation of these tools when used retrospectively, as they rely on patients to accurately recall their preoperative symptoms. A database search of Cochrane Library, PubMed, Medline Ovid, and Scopus until May 2021 was completed in duplicate by two reviewers. Studies considered eligible for inclusion were those which reported on patient recall bias associated with orthopedic surgery. The primary outcome of interest investigated was the accuracy of patient recollection of preoperative health status. Any factors that were identified as affecting patient recall were secondary outcomes of interest. Of the 4,065 studies initially screened, 20 studies with 3,454 patients were included in the final analysis. Overall, there were 2,371 (69%) knee and hip patients, 422 (12%) shoulder patients, 370 (11%) spine patients, 208 (6%) other upper extremity patients, and 83 (2%) foot and ankle patients. Out of the eight studies that evaluated patient recall within three months postoperatively, seven studies concluded that patient recall is accurate. Out of the 13 studies that evaluated patient recall beyond three months postoperatively, nine studies concluded that patient recall is inaccurate. The accuracy of patient recall of preoperative symptoms after elective orthopedic procedures is not reliable beyond three months postoperatively.

Are the results of patient reported outcome measures after spine surgery influenced by recall of preoperative scores? A randomized controlled trial

BACKGROUND CONTEXT

Patient reported outcome measures (PROMs) are of utmost importance to clinical practice as they permit a patient-focused evaluation of surgical outcomes. However, recall bias can limit an adequate interpretation of PROMs.

PURPOSE

To assess the impact of recall bias of preoperative status on postoperative PROMs of patients submitted to surgery due to degenerative spine disease.

STUDY DESIGN / SETTING

Randomized controlled trial in a tertiary care neurosurgical unit in Portugal PATIENT SAMPLE: All patients submitted to surgery at our institution from January 2019 to April 2020 due to degenerative lumbar or cervical spine disease with valid PROMs questionnaires were enrolled, and 2 computer generated randomized groups were created.

OUTCOME MEASURES

The study´s primary endpoint was the median postoperative Core Outcome Measure Index (COMI) score.

METHODS

The intervention group was sent postoperative questionnaires including preoperative answers, while patients in the control group were sent the same PROMs without the preoperative answers.

RESULTS

Randomization was applied to 236 patients (118 for each group) and valid results were obtained for 147 patients (81 lumbar, 44 from the intervention group; and 66 cervical, 29 from the intervention group), from which 88 (60%) were females, with a median age of 58 years. Both groups shared similar baseline clinical characteristics and preoperative scores. Median postoperative COMI scores and interquartile ranges (IQR) were 4.20 (IQR: 2.30-6.00) and 5.45 (IQR: 3.75-7.40) for the intervention and control groups, respectively (Wilcoxon, p=.02). This difference was reached mainly due to cervical spine patients as median postoperative COMI score was 3.95 (IQR: 2.20-5.32) in the intervention group and 5.1 (IQR: 4.0-8.4) in the control group (Wilcoxon, p=.01). No significant difference was reached for lumbar patients.

CONCLUSIONS

Better PROMs scores were obtained for degenerative cervical spine patients to whom the preoperative results were provided. Therefore, providing preoperative scores to patients upon postoperative PROMs fulfilment might influence postoperative results. Further research is necessary to increase the reliability of PROMs in clinical practice.

Real-Time Measurement of Patient Reported Outcomes and Opioid Use Following Urologic Procedures using Automated Text Messaging

OBJECTIVE

To evaluate opioid consumption and patient-reported pain intensity following urologic procedures.

METHODS

Adult patients were consented following a urologic procedure, and data was collected through post-operative day 28 in a large tertiary care academic health system. An automated text messaging platform was used to collect patient reported pain intensity, ability to manage pain, and opioid use measured in oxycodone 5 mg tablet equivalents. Outcomes were weighted based on the inverse probability of response to yield representative estimates.

RESULTS

1015 (51.8%) patients responded to the text-message survey. The median number of pills prescribed was 10 (IQR 6-10), and the median number of pills taken was 2 (IQR 0-6). By postoperative day 7, the median tablets taken overall was 0. Over the study period, 60.1% (6566) of all tablets prescribed were left unused, and 38.4% of patients did not use any of the prescribed opioids. Across urologic procedures, 6 tablets would accommodate the 75^th percentile of patient-reported use, with the exception of major open procedures.

CONCLUSIONS

In this study utilizing real-time measurement of opioid use and pain levels with text messaging, there was evidence of dramatic over-prescription of opioids relative to use and pain levels. Patient-reported data, collected via text messaging, can support clinicians and policy leaders in forming national guidelines on evidence-based best practices, personalizing prescriptions and guide shared decision making to decrease opioid excess.

Long-Term Reliability and Validity of the AO Spine PROST (Patient-Reported Outcome Spine Trauma)

STUDY DESIGN

Cross-sectional validation study.

OBJECTIVE

The aim was to validate the AO Spine Patient-Reported Outcome Spine Trauma (PROST) at a minimum of 12 months posttrauma and to evaluate patient characteristics, types of spine fractures, and treatment strategies as determinants of AO Spine PROST scores.

SUMMARY OF BACKGROUND DATA

The reliability and validity of the AO Spine PROST as a measure of health-related quality of life for more than 12 months after onset of spine trauma is unclear.

MATERIALS AND METHODS

Patients with a traumatic spine injury were recruited from a level-1 trauma center. They were asked to complete the AO Spine PROST, EuroQoL 5D-5L (EQ-5D-5L), and either Oswestry disability index (ODI) or neck disability index (NDI) for concurrent validity. Internal consistency was assessed by calculating the Cronbach α and item-total correlation coefficients. Test-retest reliability was evaluated using intraclass correlation coefficients. Spearman correlation tests were performed for the AO Spine PROST in correlation with the EQ-5D-5L, and either ODI or NDI. Determinants for AO Spine PROST score were analyzed using multivariate regression models.

RESULTS

A total of 175 patients participated in the cross-sectional arm and 49 in the test-retest arm of the study. Median duration of follow-up was 94.5 months. No floor or ceiling effects were seen. Internal consistency was excellent (α=0.98, item-total correlation coefficient: 0.73-0.91) as well as test-retest reliability (intraclass correlation coefficient=0.81). Satisfactory correlations were seen for the EQ-5D-5L (0.76; P <0.001), ODI (0.69; P <0.001), and NDI (0.68; P <0.001) with the AO Spine PROST. Multivariate linear regression models showed that having ≥1 comorbidities, duration of return to work within the range of 7 to 43 months and no return to work were significant independent determinants for a worse AO Spine PROST score.

CONCLUSIONS

Very good long-term reliability and validity results were found for the AO Spine PROST.

Lumbar intervertebral disc degeneration, end-plates and paraspinal muscle changes in children and adolescents with low-back pain

The prevalence of low-back pain (LBP) in adolescents ranges from 7 to 72%. We aimed to define the radiologic characteristics of the lumbar spine in children and adolescents with LBP with/without leg pain. Two hundred and fourteen children and adolescents, who were born between 2001 and 2009 and had lumbar spine MRI for LBP with/without leg pain, were evaluated in terms of intervertebral disc degeneration (IVDD), end-plates and paraspinal muscle changes on lumbar spine MRIs. Severe IVDD was detected at all lumbar levels except for L2-L3. Modic changes were present in 4.2% of the patients. Modic changes were more common in patients with severe IVDD than in those with mild-to-moderate IVDD. Severe IVDD was significantly associated with Modic changes at the corresponding L1-L2 and L3-L4 disc levels. Girls had significantly more fatty infiltration in the paraspinal muscles when compared to boys. The risk of having severe IVDD concomitant with Modic changes was high [odds ratio (OR), 8.6]. The OR was 20.7 for predicting the presence of severe IVDD at any level if Modic changes presented particularly at the L3-L4 level. The ORs of Modic changes presented at any lumbar level at the background of fat-infiltrated multifidus at L3-L4 and L4-L5 levels were 8.3 and 9.1, respectively. Fatty infiltration in the paraspinal muscles and IVDD were closely associated with Modic changes in children and adolescents with LBP. Lumbar IVDD in children and adolescents could be the result of a mechanical pathology.

A Narrative Literature Review of Bias in Collecting Patient Reported Outcomes Measures (PROMs)

There is a growing interest in the collection and use of patient reported outcomes because they not only provide clinicians with crucial information, but can also be used for economic evaluation and enable public health decisions. During the collection phase of PROMs, there are several factors that can potentially bias the analysis of PROM data. It is crucial that the collected data are reliable and comparable. The aim of this paper was to analyze the type of bias that have already been taken into consideration in the literature. A literature review was conducted by the authors searching on PubMed database, after the selection process, 24 studies were included in this review, mostly regarding orthopedics. Seven types of bias were identified: Non-response bias, collection method related bias, fatigue bias, timing bias, language bias, proxy response bias, and recall bias. Regarding fatigue bias and timing bias, only one study was found; for non-response bias, collection mode related bias, and recall bias, no agreement was found between studies. For these reasons, further research on this subject is needed in order to assess each bias type in relation to each medical specialty, and therefore find correction methods for reliable and comparable data for analysis.