A primary goal of surgical management of adolescent idiopathic scoliosis is to halt curve progression. Patients are generally considered candidates for surgical treatment when the curve magnitude reaches 45° to 50° and is likely to progress. Although this is a common threshold accepted by many surgeons, it is often difficult to gain access to the operating room in a timely manner in publicly funded health-care systems. In these systems, waiting lists are a common way of rationing scarce health-care resources.
Studies in several disciplines have investigated the impact of surgical waiting times on surgical outcomes and health-related quality of life1-5. Studies in a population of patients with scoliosis have noted curve progression during the wait for surgery. Ahn et al. studied a series of patients with heterogeneous scoliosis curve types; 42% of the patients who waited longer than six months for surgery had >10° of curve progression during the wait compared with 13% of the patients who waited less than six months; this difference was significant1. Larger curve magnitude has also been associated with increased difficulty of the surgery and with increased odds of adverse events1,6. Although these studies imply that operating on smaller curves is associated with less complexity and less potential for complications, the impact of treating larger curves on perioperative health-care resources has not been not well defined.
Since prolonged delays before surgery can lead to increasing deformity and since the surgical waiting time for patients with adolescent idiopathic scoliosis can exceed one year, quantifying the effects of treating larger adolescent idiopathic scoliosis curves is important. The purpose of this study was to quantify the relationship between the utilization of perioperative health-care resources and curve magnitude for patients with adolescent idiopathic scoliosis treated with spinal instrumentation and fusion.
Patients
To minimize confounding patient and treatment factors that would influence perioperative resource utilization, we sought to investigate our research question in a homogeneous population of otherwise healthy patients with adolescent idiopathic scoliosis who had an isolated main thoracic curve. Consequently, we used the Harms Study Group registry to identify a nested cohort of patients with Lenke7 type 1A and 1B adolescent idiopathic scoliosis curves treated between 2004 and 2009. The Harms Study Group registry is a prospective longitudinal international database of patients between the ages of ten and twenty-one years with a diagnosis of adolescent idiopathic scoliosis requiring surgical management. We further restricted our study cohort to patients with a main thoracic curve magnitude of at least 45° treated with posterior spinal instrumentation and fusion. Institutional review board approval was obtained for both the registry and this particular query of the registry database.
Resource Utilization Outcomes
To quantify the association between main thoracic curve magnitude and the utilization of health-care resources, we analyzed four perioperative resources important in all health-care systems: (1) operative time, (2) duration of hospitalization, (3) number of vertebral levels instrumented, and (4) the occurrence and quantity of allogeneic blood transfusion. These four perioperative outcomes defined the primary health-care resource outcomes of interest and were extracted from the registry for all participants.
The primary predictor of interest was curve magnitude, defined as the magnitude of the structural main thoracic scoliosis on posteroanterior radiographs. Curve magnitude and other patient and perioperative variables for each subject were also extracted from the Harms Study Group registry. A priori, we hypothesized that the following prospectively collected variables were associated with utilization of one or more of the four resource outcomes of interest: age, sex, body mass index (BMI), curve magnitude, surgical center, bone graft method, Lenke classification lumbar spine modifier, Lenke classification sagittal-thoracic modifier, lowest instrumented vertebra, instrumentation construct type, curve correction percentage, and estimated blood loss.
Statistical Analysis
Descriptive statistics were calculated for all study variables. Continuous data were summarized with use of the mean and 95% confidence interval (CI) or the median and interquartile range (IQR). Nominal and ordinal data were summarized with use of counts and proportions.
Univariate analysis was used to explore associations between the primary health-care resource outcomes and the a priori defined patient and perioperative variables. A p value of <0.05 was considered significant. Significant variables were then considered as potential predictors in subsequent multivariate regression models.
Multivariate regression models were constructed for each perioperative health-care resource outcome of interest. Linear models were used for operative time, duration of hospitalization, and number of vertebral levels instrumented; logistic regression was used to estimate the odds of receiving an allogeneic blood transfusion. The adjusted R2 statistic was used to assess the overall explanatory ability of the linear models. The curve magnitude β coefficient (slope) from each regression model was used to estimate the effect of increasing curve magnitude on resource utilization when controlling for all other included predictors.
Source of Funding
No funding was received for this study; however, the Harms Study Group receives financial support from DePuy Spine.
The results of our multicenter analysis of 325 Lenke type 1A and 1B idiopathic scoliosis curves suggest that larger curve magnitude is associated with increased perioperative health-care resource utilization. Specifically, when controlling for important variables such as surgical center, increases in curve magnitude were associated with additional operative time, number of vertebral levels instrumented, and odds of receiving an allogeneic blood transfusion. The duration of hospitalization was not associated with curve magnitude and was primarily associated with the surgical center.
These findings are consistent with previous research suggesting that treating larger curves involves greater perioperative complexity and more extensive surgery. Ahn et al. found that 14.7% of patients who waited longer than six months for surgery required additional surgery compared with the original plan at the time that surgery was first indicated1. Only 1.6% of children in the group who waited less than six months required additional surgery. Similarly, Miyanji et al. demonstrated that surgeons predicted increased complexity, operative time, blood loss, and duration of hospitalization when treating larger curves6.
Our study extends these prior reports by quantifying the effect of increasing curve magnitude on important perioperative health-care resources. The strengths of the current study include the prospective multicenter collection of data, the large sample size, and inclusion of a homogeneous curve population. Additionally, the use of multivariate regression modeling controlled for differences in patient or surgeon factors while estimating the effect of curve magnitude on the resource outcomes of interest. Furthermore, the regression models demonstrated high explanatory ability, with at least 65% of outcome variance explained by the variables included (as indicated by the adjusted R2 statistic).
Although the results demonstrated that treating larger curves was strongly associated with increased operative time, number of vertebral levels instrumented, and odds of receiving an allogeneic blood transfusion, the magnitude of increased perioperative resource utilization varied among the outcomes of interest. For example, a 10° increase in curve magnitude was associated with an additional 7.8 minutes of operative time (p = 0.03) but only an additional 0.3 level instrumented (p = 0.0005).
The clinical importance of these findings must be considered in the context of the study design. Only Lenke 1A and 1B curve types were included, which by definition represent structural main thoracic curves with nonstructural upper thoracic or lumbar components. Consequently, isolated instrumentation and fusion of the main thoracic curve is the preferred treatment. Our results supported the expectation that surgeons instrument additional levels when treating larger curves; however, the number of additional levels in our cohort was likely limited by the inclusion of only structural main thoracic curves.
Despite the positive findings of the current study, certain limitations warrant discussion. The homogeneous study population ensures the validity of the results for Lenke 1A and 1B curves; however, we are unable to ensure that these results apply to other curve types. In addition, this analysis does not attempt to quantify the increased costs associated with the health-care resources of interest. Wide variations in costs for operative time, blood products, and surgical implants exist across North America; thus, although cost estimates are of interest, providing them is often problematic and misleading. Additionally, this study did not seek to determine whether larger curves were associated with an increased rate of perioperative complications. Further research to address this question is warranted given the resulting potential for increased patient morbidity as well as resource burden. Finally, this study was not designed to answer the important question “Should surgeons operate on smaller curves to prevent potential problems?” However, we believe that the relationship between increasing curve magnitude and increasing resource utilization does need to be described to inform physicians, patients, and health-care policy makers.
In conclusion, lengthy surgical waiting lists have been reported to be associated with increases in scoliosis curve magnitude and in the perceived difficulty of treating these curves. The present study of a large multicenter patient cohort demonstrated increases in operative time, number of levels instrumented, and odds of receiving an allogeneic blood transfusion when treating larger Lenke 1A and 1B curves. As difficult funding decisions for health care continue to be made, physicians, patients, and health-care policy makers must recognize that treatment of larger curves is associated with greater use of perioperative health-care resources compared with smaller curves.
Disclosure: One or more of the authors received payments or services, either directly or indirectly (i.e., via his or her institution), from a third party in support of an aspect of this work. In addition, one or more of the authors, or his or her institution, has had a financial relationship, in the thirty-six months prior to submission of this work, with an entity in the biomedical arena that could be perceived to influence or have the potential to influence what is written in this work. No author has had any other relationships, or has engaged in any other activities, that could be perceived to influence or have the potential to influence what is written in this work. The complete Disclosures of Potential Conflicts of Interest submitted by authors are always provided with the online version of the article.