Rates and Predictors of Surgery for Lumbar Disc Herniation Between the Military and Civilian Health Care Systems

ABSTRACT

Study Design

Retrospective review (level of evidence III).

Objective

Surgical care patterns for lumbar disc herniation (LDH), a common musculoskeletal condition of high relevance to the Military Health System (MHS), have not been described or compared across the direct care and purchased care MHS components. This study aimed to describe surgery rates in MHS beneficiaries who were diagnosed with LDH in direct care versus purchased care and to evaluate characteristics associated with the location of surgery. Differences in care patterns for LDH may suggest unexpected variation within the centrally managed MHS.

Methods

We described 1-year rates of surgery among beneficiaries who were diagnosed with LDH in direct care versus purchased care. Among beneficiaries who were diagnosed in direct care and had surgery, multivariable logistic regression models were used to identify characteristics associated with surgery location.

Results

We identified 726,638 MHS beneficiaries who were diagnosed with LDH in direct care or purchased care during the 9-year study period. One-year surgery rates were 10.1% in beneficiaries who were diagnosed in direct care versus 11.3% in beneficiaries who were diagnosed in purchased care. Among the 7467 patients who were diagnosed in direct care and had surgery within 1 year, characteristics associated with lower probability of surgery in purchased care versus direct care included diagnosing facility type (hospital with a neurosurgery or spine specialty versus clinic (odds ratio [OR], 0.12 (95% CI, 0.10-0.15)), Navy versus Army (OR, 0.24 (95% CI, 0.21-0.28)), and diagnosing facility specialty (Medical Expense and Performance Reporting System) (surgical care (OR, 0.33 (95% CI, 0.27-0.40)) and orthopedic care (OR, 0.39 (95% CI, 0.33-0.46)) versus primary care. The presence of comorbidities was associated with higher probability of surgery in purchased care versus direct care (OR, 1.20 (95% CI, 1.06-1.36)).

Conclusions

The 1-year rate of surgery for LDH was modestly higher in beneficiaries who were diagnosed in purchased care versus direct care. Among patients who were diagnosed in direct care, several patient-level and facility-level characteristics were associated with receiving surgery in purchased care, suggesting potentially unexpected variation in care utilization across components of the MHS.

INTRODUCTION

Lumbar disc herniation (LDH) is a common musculoskeletal pathology with symptomatic LDH effecting 1% to 2% of the US population^1,2 as well as a cause of low back pain, a highly prevalent cause of disability in the US civilian and military populations.¹ US military service members are highly active and have physically demanding job requirements that increase susceptibility to musculoskeletal injuries, a class of injuries recognized as a major challenge to readiness for military deployment. We previously characterized risk factors associated with progression to surgical intervention for LDH in Military Health System (MHS) beneficiaries after diagnosis in military treatment facilities (MTFs).² Patterns of surgical care for LDH in the MHS, however, have not been described. Potential unexplained heterogeneity in LDH surgical care rates within the MHS may suggest unexpected variation in care utilization in the centrally managed MHS.

The MHS provides health care to approximately 9.5 million active duty and retired military personnel and their civilian dependents through a “direct care” system of MTFs staffed and operated by military employees or through care purchased from civilian health care providers (“purchased care” system). MHS beneficiaries generally can be seen in both the direct care and civilian care systems for most health care services at little or no cost to the beneficiary.³ Active duty beneficiaries have priority access to MTF care, and most active duty families select a MTF primary care provider. However, “leakage” between health systems may occur, for example, when an MTF does not have capability or capacity to provide required specialty care within a specified time frame.³

Several studies have evaluated geographic variation in care as a proxy for efficiency in the MHS,³ with lower care utilization described for hospital referral regions with versus without a MTF.⁴ Rates and location of surgery for LDH, a common procedure of high relevance to the military population, have not been characterized across MHS components. We sought to compare surgical care patterns in MHS beneficiaries who were diagnosed with LDH and treated in the direct care versus purchased care system.

METHODS

Guidelines/Data Source

The Military Data Repository (MDR) contains comprehensive records of all health care encounters occurring within the MHS for approximately 9.5 million TRICARE Health Plan beneficiaries including active duty service members, dependents, retirees, and those with medical disabilities. The MDR includes person-level or claims-level records from health care services for which TRICARE is the payer in either MTFs or civilian facilities. This retrospective, observational study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) for reporting both an accurate and complete observational study.⁵

Study Design

After institutional review board approval, the MDR was queried for all patients with the most commonly used International Classification of Diseases (ICD) diagnostic code for LDH (M5126, 72210) from FY2010-2018; these codes are used to classify invertebral disc displacement in the lumbar region. The earliest such diagnosis in an outpatient encounter in direct care or purchased care was identified for each patient as the initial diagnosis. Among this cohort, we identified patients who were diagnosed with LDH in direct care and had surgical intervention in either direct care or in purchased care within 1 year postdiagnosis, a clinically relevant duration of follow-up for conservative management of LDH. Patients were excluded from analysis if they were diagnosed in a facility other than a clinic or hospital (N = 3944), if the diagnosing facility Medical Expense and Performance Reporting System (MEPRS) service code was not classified within codes for primary care, surgical care, orthopedic care, emergency medicine, and rehabilitation care services (N = 1898), and if patients were missing one of the included covariates (N = 213).

Variables

Surgical intervention for LDH was defined as a Current Procedural Terminology code for lumbar microdiscectomy (63030, 63035, and 62380) or lumbar decompression (63005, 63012, 63017, 63042, 63044, 63047, 63048, 63056, and 63057). Patient characteristics included age at the time of diagnosis, gender, history of self-reported tobacco use (yes/no) at any time during the study period, presence of health comorbidities (yes/no), and beneficiary category (active duty, retired, and all others). Presence of comorbidities was defined as diagnosis over follow-up of one or more chronic condition classes identified in the MDH health risk file based on relevant ICD codes: cancer, cardiovascular disease, cerebrovascular disease, diabetes, osteoarthritis, rheumatoid arthritis/inflammatory, renal, and respiratory disease. Treatment and provider characteristics included diagnosis facility type (“hospital” includes sites with admitting/inpatient capabilities, whereas “clinic” includes MTFs with outpatient capabilities only), diagnosing facility (Army, Navy, Air Force, and Defense Health Agency (DHA)/joint facility), MEPRS medical service code (primary care, orthopedic care, surgical care, emergency medicine, and rehabilitation), and presence of fellowship trained spine surgeons (yes/no).

Outcome of Interest

The primary outcomes of interest were rates and predictors of surgical intervention for LDH in direct care versus purchased care within 1 year postdiagnosis.

Statistical Analysis

Among MHS patients diagnosed with LDH, we described and compared patients who had surgery within  year when diagnosed in direct care versus purchased care. Denominators excluded patients (N = 36,365 [5.0%]) without reported surgery who did not remain MHS eligible through a 1-year follow-up. Logistic regression was used to estimate the association of diagnosis location with the probability of surgery within 1 year, with adjustment for age, sex, and active duty status; these relevant demographic characteristics are reliable and available in the MDR for both direct care and purchased care encounters.

In the subgroup of beneficiaries who were diagnosed with LDH in direct care and had surgery within 1 year postdiagnosis, we used logistic regression to estimate unadjusted and adjusted associations (odds ratios [ORs]) of patient and diagnosing facility and service characteristics as independent variables and surgical intervention in purchased care versus direct care as the dependent variable. Presence of patient comorbidities was derived from diagnosis codes and pharmacy records. Specialty hospitals were defined by the presence of spine surgical training programs (orthopedics or neurosurgery) with spine surgery billets during the study period. The model was repeated restricting to active duty service members to evaluate for potential qualitative differences in referral patterns compared with non–active duty beneficiaries. Self-reported race was included as a covariate only in the active duty analysis because race is frequently not reported (>30% missingness) for beneficiary family members.⁶

All reported P values are two-sided with the significance level set to <0.05. All analyses were performed using R (Version 3.6.1 Vienna, Austria).

RESULTS

Patient Population

We identified 726,638 patients in the MHS who were diagnosed with LDH during the study period in direct care or purchased care. One-year rates of surgery were 10.1% among patients who were diagnosed in direct care versus 11.3% among patients who were diagnosed in purchased care. Patients who were diagnosed in direct care versus purchased care were younger (mean age [SD] = 39.4 [11.5] versus 59.8 [16.9] years) and were more likely to be male (74.9% versus 46.8%) and active duty members (71.9% versus 12.3%). In logistic regression models adjusted for age, sex, and active duty status, diagnosis in purchased care versus direct care was associated with higher probability of surgery: OR, 1.20; 95% CI, 1.16-1.23; P < 0.001.

Surgical Intervention at a MTF versus Civilian Center

Among the subgroup of patients who were diagnosed in direct care and had surgery within 1 year (N = 7467), several diagnosing facility features and provider features were associated with higher probability of surgery in purchased care versus direct care; these included clinic versus hospital with a neurosurgery or spine specialty; Army versus Navy facility; and primary care versus surgical, orthopedic, and rehabilitation care MEPRS (Table I). Among patient characteristics, older age and the presence of comorbidities were each associated with higher and active duty status associated with lower probability of surgery in purchased care. The OR associated with presence of comorbidities was 1.20 (95% CI, 1.06-1.36; P < 0.01) with adjustment for patient and facility characteristics (Table II).

Active Duty Service Members Surgical Intervention at a MTF versus Civilian Center

In analyses repeated among active duty service members only (N = 3204), associations did not substantially change for characteristics other than surgical care MEPRS and comorbidities, the ORs for which were attenuated and no longer significant. Self-reported race, age, and sex were each not associated with location of surgery (Table III).

DISCUSSION

We compared surgical care patterns for LDH in the MHS direct care and purchased systems. We found comparable 1-year rates of surgery in those diagnosed with LDH in direct care (10.1%) versus purchased care (11.3%). Among patients diagnosed in direct care, we identified several facility and patient-related characteristics, including presence of comorbidities, associated with higher probability of surgery in purchased versus direct care. Similar to our findings, previous national health care system–based registry analyses have reported incidence and outcomes of surgical intervention in patients diagnosed with LDH.⁷ Differences in surgical care patterns for LDH, a common condition of high relevance to the MHS population, may suggest unexpected variation within the centrally managed MHS.

Health risk factors in LDH patients have previously been found associated with increased risks of surgical intervention and poor surgical outcomes including greater blood loss, operative times, and prolonged hospitalization.^{8 89,109,10} We found, perhaps surprisingly, that the presence of comorbidities in MHS beneficiaries who were diagnosed with LDH in direct care was associated with higher probability of surgery in purchased versus direct care, independent of the presence of spine surgical subspecialists at the diagnosing MTF. This finding, however, was attenuated in a subanalysis of active duty service members, suggesting that this potentially unexpected difference in surgical care patterns for LDH was more pronounced among retirees and dependents than the generally younger and healthier active duty population.

Among the full cohort of beneficiaries diagnosed with LDH across the MHS, the modestly higher surgery rate in beneficiaries who were diagnosed in purchased versus direct care was in part explained by key demographic differences in the patient populations served by these care settings. Active duty service members were less frequently diagnosed in purchased care than in the direct care system that is designed to facilitate their access to care. However, of the active duty service members who were diagnosed in a MTF and had surgery, nearly 40% had surgery in purchased care. This substantial referral rate may reflect lack of capability or capacity within the required time frame at the diagnosing MTF.³ Among beneficiaries who were diagnosed at an MTF, we found that the presence of a surgical spine subspecialist was associated with a higher probability of surgery in a MTF versus purchased care. This finding is unsurprising, as the decision to seek surgery in either setting may be largely influenced by proximity or ease of access to care.

This study has several limitations. Analyses of health care databases are inherently challenged by risk of coding errors and differential coding across facilities and limited available patient-level characteristics. Additionally, our analyses did not evaluate geographic characteristics of individual MTF regions and therefore could not distinguish the importance of proximity to the diagnosing MTF. Separate analyses of care utilization metrics are warranted to directly evaluate geographic variation in spine surgical care utilization across MHS components.

This study is the first to our knowledge to characterize rates of surgical care for LDH across components of the MHS. We identified several patient- and facility-level characteristics associated with the location of surgery for this common spine procedure of high relevance to the MHS. More work is needed to identify underlying causes of these potentially unexplained differences in surgery location and their cost impact in the MHS.

ACKNOWLEDGMENTS

The authors would like to acknowledge the contributions of their colleague.

FUNDING

None declared.

CONFLICT OF INTEREST STATEMENT

None declared.

CLINICAL TRIAL REGISTRATION: IDENTIFIER

None declared.

INSTITUTIONAL REVIEW BOARD (HUMAN SUBJECTS)

This study was approved by the Walter Reed National Military Medical Center.

INSTITUTIONAL ANIMAL CARE AND USE COMMITTEE (IACUC)

None declared.

INDIVIDUAL AUTHOR CONTRIBUTION STATEMENT

N.L.W., D.I.B., A.B.A., S.C.W., collected and analyzed the data and drafted the original manuscript. C.J.N. collected data and reviewed the manuscript. All authors read and approved the final manuscript.

DATA AVAILABILITY

The data that support the findings of this study are available on request from the corresponding author.

INSTITUTIONAL CLEARANCE

None declared.

REFERENCES

Author notes