Keywords
scaphoid fracture - insurance - scaphoid nonunion
Appropriate initial treatment of scaphoid fractures routinely leads to reliable healing.[1]
[2] Scaphoid nonunion, on the other hand, may occur in up to 10% of scaphoid fractures.[3] When left untreated, nonunion of the scaphoid results in radiocarpal arthrosis[4] and requires challenging surgical solutions.[5] Several factors have been associated with nonunion development, including delayed
diagnosis, proximal fracture location, degree of fracture displacement, and carpal
instability.[6]
Many factors may play into the timing of diagnosis and treatment of scaphoid fractures.
The purpose of our study was to evaluate whether or not insurance status plays a role
in how patients with scaphoid fractures are initially managed. Specifically, we hypothesized
that a disproportionate percentage of underinsured individuals present with symptomatic
scaphoid nonunion relative to those presenting with primary scaphoid fracture.
Methods
After obtaining Institutional Review Board approval for this study, we performed a
retrospective case–control study on patients who presented to a single surgeon at
a tertiary referral center between August 2006 and March 2015 with any of the following:
new patient or consult evaluation for an (International Classification of Diseases
(ICD)-9 diagnosis code 814.01 (scaphoid fracture) and/or 733.82 (fracture nonunion).
In addition, using Current Procedural Terminology (CPT) code 25628 (primary scaphoid
fracture repair) and CPT code 25540 (scaphoid nonunion repair), we searched for patients
who had undergone surgical treatment for scaphoid fractures or scaphoid fracture nonunions.
This study was performed in the United States.
After reviewing patient radiographic imaging, we defined cases as patients presenting
with scaphoid nonunions and controls as patients presenting with primary scaphoid
fractures. We excluded patients with other concomitant injuries such as distal radius
or elbow fractures. We confirmed the appropriate diagnosis by chart review and image
review, and not solely based on the ICD-9 diagnosis code or CPT code. We characterized
patients as underinsured or insured. Underinsured patients had no medical insurance
or Medicaid/state-funded insurance. Insured patients had Medicare or private insurance.
We collected demographic information on each patient, including age, gender, and fracture
displacement (>1 mm) based on computed tomography (CT) scan, fracture location, laterality
of fracture, and insurance status by retrospectively reviewing patient medical records
(both paper and electronic) and radiographs. Imaging measurements were performed by
the surgeon and two senior orthopaedic residents using Synapse software (Fujifilm).
Fractures were classified according to the Mayo system to determine proximal and distal
pole fractures.[7]
Statistical analyses were conducted in Stata version 13.1 (StataCorp LP, College Station,
TX). We used Pearson's chi-square test or Fischer's exact test to determine whether
cases and controls differed in terms of age, gender, fracture displacement at presentation,
fracture location, laterality of fracture, and insurance status. We then conducted
a multivariate analysis using a backward stepwise process in which all variables associated,
in bivariate analyses, with case/control status were initially included, and variables
with the highest p-values were sequentially eliminated until all remaining variables had p < 0.20. The level of significance was set at p < 0.05.
Results
A total of 71 patients were identified. Of these, 32 (45%) were primary fractures
(controls) and 39 (55%) were nonunions (cases). Nonunion patients presented on average
205 days after injury. Acute fracture patients presented on average 30 days after
injury. Nonunions patients were more likely than acute fracture patients to have had
fracture displacement at the time of presentation to our center (72 vs. 41%; p = 0.015; [Table 1]). This finding makes sense, as nonunions, by nature, have a gap between the bone
fragments and often a humpback deformity with displacement of the distal fragment.
Nonunions were also more likely than controls to have proximal pole fractures (18
vs. 0%; p < 0.001) and less likely than controls to have fractures located at the distal aspect
(0 vs. 19%; p < 0.001).
Table 1
Characteristics of patients presenting with primary scaphoid fractures (controls)
versus those presenting with scaphoid nonunions (cases)
|
Primary fracture (controls, N = 32)
|
Nonunion (cases, N = 39)
|
p-Value
|
|
Age (years ± SD)
|
29.7 ± 19.1
|
25.7 ± 13.5
|
0.302
|
|
Male
|
25 (78%)
|
33 (85%)
|
0.547
|
|
Female
|
7 (22%)
|
6 (15%)
|
|
Nondisplaced fracture
|
19 (59%)
|
11 (28%)
|
0.015
|
|
Displaced fracture
|
13 (41%)
|
28 (72%)
|
|
Proximal pole
|
0 (0%)
|
7 (18%)
|
<0.001
|
|
Mid/waist
|
26 (81%)
|
32 (82%)
|
|
Distal pole
|
6 (19%)
|
0 (0%)
|
|
Left
|
13 (41%)
|
17 (44%)
|
0.815
|
|
Right
|
19 (59%)
|
22 (56%)
|
|
Insured
|
26 (81%)
|
21 (54%)
|
0.023
|
|
Underinsured
|
6 (19%)
|
18 (46%)
|
Abbreviation: SD, standard deviation.
Finally, nonunions were more likely than controls to be underinsured (46 vs. 19%;
p = 0.023). In the final multivariate model, the only factor that remained statistically
associated with case/control status was underinsurance: nonunions were more likely
than controls to be underinsured (odds ratio = 3.7; 95% confidence interval = 1.3–11;
[Fig. 1]). To provide a context of the percentage of underinsured in the community, the treating
surgeon's breakdown, using the same definition of insured and underinsured patients
in 2011 (a midpoint in the study collection), was approximately 85% insured and 15%
underinsured.
Fig. 1 This bar graph demonstrates the percentage difference of insurance status between
the primary fracture group and the nonunion group. Nonunions were more likely to be
underinsured than controls (odds ratio = 3.7; 95% confidence interval = 1.3–11).
We were not able to obtain complete demographic data regarding smoking, diabetes,
and immunocompromised status in 12 patients in the control group and 10 patients in
the nonunion group. Two patients in the control group and five patients in the nonunion
group were current smokers. None of the patients in the nonunion group had diabetes;
one patient was immunocompromised secondary to hepatitis C.
Discussion
These results confirmed our hypothesis that patients presenting with scaphoid nonunion
are more likely to be underinsured than patients presenting with primary scaphoid
fractures. This discovery suggests that underinsurance is a risk factor for the development
of nonunion. It is important to recognize that this study implies an association rather
than causation. Potential explanations for these results include disparities in time
to treatment, inadequate treatment, patient noncompliance, and patient comorbidities
such as smoking, diabetes, and hepatitis C between the two groups. Our study is limited
by the fact that we were unable to obtain these specific historical data of all of
our patients to create a more robust assessment of causative factors.
Due to brief and inconsistent history collections in many of our patients' charts,
we are unable to offer comprehensive preoperative information on the nonunion patient
cohort, specifically regarding the patients' own explanation of obstacles to treatment.
However, these patients did describe two common scenarios that may have predisposed
the underinsured group to nonunion: delay in diagnosis and inadequate fracture management.
For example, some patients never sought treatment after their injury, assuming they
sprained their wrist. Other patients did present to a medical facility for initial
management of their injury. Emergency or urgent care centers would diagnose a sprain,
supply a splint or brace, and suggest follow-up with a hand surgeon. Some patients
may have seen a treating physician at one point in their initial management but did
not complete follow-up care due to their insurance or lack thereof. A few patients
described treating physicians telling them that their scaphoid fractures had healed
after 4 to 6 weeks and to return as needed. Our study is also limited by the fact
that we cannot honestly report these additional clinical characteristics in a scientific
fashion.
In conjunction with the passage of the Affordable Care Act in 2010, there has been
an increased focus on healthcare disparities for the underinsured/noninsured. Studies
have demonstrated that insurance status influences orthopaedic outcomes such as timely
access to specialized care[8]
[9] and continuity of care.[10] The lack of timely access by the underinsured has significant implications for scaphoid
fracture care, in particular, as treatment delays alter the course of healing.
Skaggs et al demonstrated that timely access to pediatric orthopaedic care for a fictitious
child with a forearm fracture was available in 100% of offices polled when the child
was said to have private insurance versus in 2% of offices for a child with California
Medicaid (Medi-Cal).[11] Patterson et al reached similar conclusions using a fictitious adult patient, with
an appointment for outpatient subspecialty care being offered within 2 weeks in 79%
of offices when the patient was said to have private insurance and only 59% of offices
for the patient with North Carolina Medicaid.[12] Interestingly, they also found that practices in more populous areas and closer
to academic medical centers were less likely to offer timely care. Given that delay
between fracture and intervention is a known risk factor for the development of nonunion,
we speculate that the association between nonunion and underinsurance is mediated
through this delay.
This study has few strengths. A single surgeon practicing at a single tertiary referral
center evaluated and treated all of the patients in this case–cohort study. We set
out to answer a concise question, namely does insurance status impact scaphoid fracture
healing. Our results revealed that underinsured patients had a higher incidence of
scaphoid nonunion upon presentation than insured patients. With this knowledge, one
then could consider underinsurance to be an indicator for the surgical management
of acute scaphoid fractures, though concern for the inability to follow up could still
deter surgeons from operating. Several studies have shown surgical repair of scaphoid
fractures using a headless compression screw to be a reliable surgical procedure without
prolonged postoperative immobilization.[1]
[13]
[14] Patients who are prone to inconsistent follow-up may be better suited with a scaphoid
screw (internal immobilization) rather than a cast (external immobilization) to diminish
the risk of scaphoid nonunion.
Limitations of this study include its retrospective nature and narrow scope. However,
the limited data on the presumed etiology of the scaphoid nonunion in the nonunion
cohort group truly limit this study's effective reporting on what causes the scaphoid
nonunion. It would be advantageous to conduct interviews with patients to determine
the specific nature of their postinjury course to better elucidate factors other than
the relationship between underinsurance and nonunion development. Most of our patients
were teenagers or young adults who do not always offer consistently reliable histories;
many were seen in a clinic rather than an office setting, with sparse history intake.
It also is difficult to arrange follow-up assessments of this group of patients, many
of whom commuted from a distance to reach our referral center. Once they were pain-free,
they typically would not return for regularly scheduled office visits.
In conclusion, patients who present with a scaphoid nonunion are significantly more
likely to be underinsured. Based on this finding, we recommend increased diligence
when managing an underinsured patient with a scaphoid fracture. We now consider underinsurance
to be reasonable indicator for the operative treatment of acute scaphoid fractures.
