Setting and data sources
This study employed a retrospective cohort design using the UK Hospital Episode Statistics (HES) database from NHS Digital and the UK Clinical Practice Research Datalink (CPRD). Approval was granted by the Independent Scientific Advisory Committee (ISAC) for Medicines and Healthcare products Regulatory Agency (ISAC Protocol 14-180R). These data and the study methodology were described in detail previously [1,5].
Study participants
Patients aged 18 and older, who underwent one or more lumbar procedures from April 2007 through to March 2012 were eligible for inclusion. Index operative procedures included any single procedure or combination of discectomy/microdiscectomy, excision of lumbar intervertebral disc, laminectomy, foraminotomy, lumbar decompression (or fenestration) or instrumented lumbar fusion (including all anterior and posterior approaches as well as combined approaches). For the purposes of cleanly comparing instrumented and non-instrumented surgery, we excluded patients who underwent non-instrumented or undetermined spinal fusion. Patients were required to have a minimum six months pre-index surgery data without evidence of prior spinal surgery and at least two years of postoperative follow up data.
Definition of Persistent Postoperative Pain
We categorized each surgery as a ‘success’ (i.e., no evidence of PPP) or ‘failure’ (i.e., evidence of PPP) depending on indications in the CPRD-HES data that the patient experienced ongoing pain, continuing past the ‘expected’ period for recovery following index lumbar surgery. The definition of six months was applied for the expected postoperative recovery time. The terms ‘success’ and ‘failure’ are used here in the sense of success or failure of the surgery to resolve or relieve pain, rather than anatomical success or failure of the surgical procedure. As some patients may initially improve following surgery before pain returns [6-8], we chose a period of 18 months (6-24 months postoperatively) to screen for evidence of ongoing or recurrent pain. Since there are no specific diagnosis codes that may be used to identify PPP, our estimates were based upon records of additional surgery or other interventions and attendance at pain clinics. The data screening criteria used to infer ongoing or recurrent pain are summarized in Table 1. Prescription of analgesics was not included in these criteria as patients may be prescribed analgesics for other non-spinal painful conditions.
Table 1
Criteria for evidence of persistent postoperative pain after index lumbar surgery
Any one or more of the following:
|
HES Inpatient
|
Any lumbar surgery: OPCS code 6–24 months post index surgery
|
HES Inpatient / HES Outpatient / CPRD Gold
|
Any surgical intervention for lumbar pain (e.g., neuromodulation, implantation of drug infusion delivery system): OPCS code or READ code at any time post index surgery
|
HES Outpatient / CPRD Gold
|
At least one pain-related GP visit or specialist pain clinic visit in each of two consecutive quarters occurring 6–24 months post index surgery
|
Healthcare costs
All costs were estimated from the perspective of the UK National Health Service (NHS). We first estimated the index surgery costs, including all costs incurred for the entire index inpatient episode of care. We then estimated total healthcare costs over the 24 month period after index hospital discharge by classifying health care encounters into major categories of health care resource utilization (i.e., inpatient hospital stays, outpatient clinic visits, accident and emergency attendances, primary care encounters, and medications prescribed in primary care). Respective unit costs were applied [9-12]. To account for inflation and variations in pricing over time, 2013 unit costs were applied to all years.
Statistical analyses
To estimate rates of PPP, we computed the number of patients who met the criteria for PPP as a percentage of all patients who underwent lumbar surgery (instrumented fusion versus decompression/discectomy) within the time frame.
We compared PPP rates, and two-year postoperative costs, of patients who underwent instrumented fusion with those who had decompression or discectomy using a control group selected with 1:1 exact matching based upon patient’s age at surgery and sex to account for any differences in likelihood of receiving instrumented surgery by age or sex. Matching was not undertaken on pre-surgical comorbid conditions. We examined the set of conditions that comprise the Charlson Comorbidity Index (CCI) evaluated in the one year period prior to surgery but these were not found to be predictive of the choice of surgical mode.
Confidence intervals (95% CIs) for healthcare costs were estimated using bootstrapping to allow for non-normality of the means. The difference in PPP between cases and controls was assessed using a chi-square test of proportions.
Finally, sensitivity analyses were conducted in which the key outcome variables, PPP and costs, were adjusted to account for pre-index surgery comorbid conditions (from the CCI) using logistic regression (for the probability of developing PPP) and a generalized linear regression model (using a log link and gamma distribution, for cost estimation).
All data manipulation and analyses were conducted using SAS software, Version 9.4 for Windows [SAS Institute, Cary NC].