Collecting and predicting patient reported outcomes in chiropractic practice
Student thesis: Doctoral Thesis
The eight refereed publications and four abstracts of presentations which form the basis of this PhD each deal with patient health outcomes. The publications are drawn predominantly from practice based research in chiropractic services. In a systematic review of the impact of Patient Reported Outcome Measures (PROMs) on the process and outcomes of care for a single patient, Paper 1 describes this occurring across four domains; patient assessment and initial clinical decision making, tracking progress and evaluating current treatment efficacy, influencing the patient / clinician relationship and there is weak evidence to suggest they directly influence patient outcomes. Paper two is a descriptive review of the utility of PROMs to include their ability to improve communication and shared decision making in the patient / clinician relationship. Care Response is a novel, free to use multilingual electronic PROM system developed by the author. It has had significant impact in the chiropractic profession in Europe and Canada and has contributed data to 11 peer reviewed papers and four post graduate degrees (Abstract 1) Electronic PROM systems suffer from lower response rates than paper based systems. Abstract 2 reports a study looking at the impact of this missing information on the generalisability of the overall data collected. Non respondents to an emailed assessment 30 days after starting care were less likely to have had >30 days pain in the last year but were not otherwise significantly different from those returning electronic assessments. In a telephone survey comparing respondents and non respondents, patients global impression of change (PGIC) scores were identical and there was no statistical difference in pain scores. Paper 3 sought to ascertain if patient less likely to do well with chiropractic care could be identified from data routinely collected at baseline in chiropractic practice. Longer duration of symptoms at presentation, females with higher social disability scores and males with more adverse scores for depression were found less likely to describe themselves as much improved a month after starting care. In investigating for a relationship between outcome and components of the fear-avoidance model for chronicity in lower back pain, paper 4 found only a week relationship with catastrophisation at baseline however patient’s scores of catastrophisation, fear avoidance beliefs and low self efficacy just before their second visit were significantly associated with a poorer outcome. Paper 5 looking into a relationship between the risk category patients were placed into by the STarT Back Tool reported that whilst HIGH risk patients has more adverse scores for pain at presentation this rapidly faded and at 30 and 90 days there was no significant difference between the risk groups and patients reports of their recovery (PGIC) . Comparing the health outcomes of 8222 patients accessing chiropractic services either via the NHS or privately Paper 6 described those accessing a NHS route to have had symptoms for longer and more adverse scores across a range of health domains at presentation and to be less well 30 & 90 days later. However both NHS and private patient groups improved well and differences between the two disappeared when controlling for differences at baseline. Paper 7 sought to explore the ability of chiropractic clinicians working from 5 linked practices to identify those patients less likely to do well with care at the time of their initial assessment. It concluded that they generally failed to reliably predict outcomes with most practitioners doing no better than chance. The STarT Back Tool is increasingly being recommended to guide decisions as to care pathway for patients with spinal pain. Paper 8 looked to see if the timing of when this assessment was made had any impact on its ability to detect groups of patients responding differently when undergoing a course of chiropractic care. In the assessed population (n=749) attending 1 of 11 clinics in the UK there was a significant difference with ranking at the time of presentation being unrelated to outcome. When repeated a few days post initial visit over one third of patients had changed risk group with this subsequent group being found to be an independent predictor of improvement in multivariate analysis.
7.76 MB, PDF document