Medicaid Payer Status Is a Predictor of Early Postoperative Pain Following Upper Extremity Procedures

Abstract

Background: The purpose of this study was to investigate the relationship between insurance status and patient-reported pain both before and after upper extremity surgical procedures. We hypothesized that patients with Medicaid payer status would report higher levels of pre- and postoperative pain and report less postoperative pain relief. Methods: In all, 376 patients who underwent upper extremity procedures by a single surgeon at an academic ambulatory surgery center were identified. Patient information, including insurance status and Visual Analog Scale pain score (VAS-pain) at baseline, 2 weeks, and 1, 3, and 6 months, were collected. VAS-pain scores were compared with t-tests and linear regression. Results: Preoperatively and at 2-week, 1-month, and 3-month follow-up, Medicaid patients reported statistically significant higher pain levels than patients with Private insurance, finding a mean adjusted increase of 0.51 preoperatively, 0.39 at 1 month, and 0.79 at 3 months. Preoperatively and at 3-month follow-up, Medicaid patients reported statistically significant higher pain than patients with Medicare, finding increases in VAS-pain of 0.99 preoperatively and 0.94 at 3 months. There was no difference in pain improvement between any insurance types at any time point (all P values > .05). Conclusions: Patients with Medicaid report higher levels of preoperative pain and early postoperative pain, but reported the same improvement in pain as patients with other types of insurance. As healthcare systems are becoming increasingly dependent on patient-reported outcomes, including pain, it is important to consider that differences may exist in subjective pain depending on insurance status.

Introduction

Healthcare disparities across socioeconomic status (SES), race, and ethnicity have been well-described.^1-4 Despite efforts to close this gap, evidence suggests differences in patient perception of healthcare, patient satisfaction, and clinical outcomes still exist.^5-9 Studies in multiple surgical fields have shown insurance status can be a predictor of poor outcomes following surgery, as patients with government-sponsored insurance experience more complications and increased mortality following surgery.^10-16 Additionally, lower SES has been linked to higher levels of patient-reported pain.^6-9 As Medicaid is primarily utilized by lower SES patients, these studies lend support to the anecdotal finding that Medicaid patients experience more preoperative pain and less pain relief following surgery than those with private insurance.

The passage of the Patient Protection and Affordable Care Act in 2010 expanded Medicaid eligibility and has resulted in a 30% increase in Medicaid enrollment. ¹⁷ As the number of patients enrolled in Medicaid increases, it is important to understand the relationship of insurance status to postoperative outcomes. A greater understanding of the factors influencing preoperative pain levels and postoperative pain relief can help to improve patient education of what to expect before and after surgery, leading to improved patient care and satisfaction. Additionally, further knowledge of the differences in characteristics among patients with Medicaid can help to guide areas in which care can be improved to minimize disparities in outcomes.

This study compares preoperative pain levels, postoperative pain levels, and pain relief between patients with varying insurance status. We hypothesized that Medicaid patients would report higher preoperative pain levels and less pain relief following surgery relative to patients with Medicare or private insurance. Additionally, we hypothesized similar results when comparing underinsured patients (Medicaid and uninsured) to well-insured patients (Medicare and private insurance) and when comparing patients with private insurance to those with Medicaid, Medicare, or uninsured.

Methods

This retrospective review of prospectively collected data was approved by the institutional review board prior to initiation of the study. We identified all patients who underwent upper extremity surgery between September 1, 2015 and July 1, 2017. All procedures were performed by the same hand and upper extremity fellowship trained surgeon at a single institution. Patients were included in this study if they underwent upper extremity surgery by the single surgeon and were between the ages of 18 to 80 years at the time of surgery. Of the 376 patients who underwent upper extremity surgery by the single physician between the inclusion dates, 357 met inclusion criteria. Nineteen patients were excluded because they were <18 years or >80 years of age.

Chart review was performed to obtain patient demographics, including age at the time of surgery, sex, race, body mass index (BMI), and insurance information. The patients’ prior treatment, comorbidities, medications, and smoking status were recorded as part of the initial intake form for new patients. Primary outcomes measured included preoperative and postoperative (2 weeks, 1 month, 3 months, and 6 months) pain scores. The 6-month scores were analyzed when available as not all patients had a 6-month follow-up. Pain was measured through the Visual Analog Scale pain score (VAS-pain), which requires patients to record a numeric pain score from 0 to 10, with 0 being no pain and 10 being the worst possible pain.

Prior to analysis, patients were grouped into cohorts. For the first set of analyses, the patients were divided into 4 cohorts: Cohort 1 included patients with Medicaid insurance (n = 106), Cohort 2 included patients with Medicare insurance (n = 85), Cohort 3 included patients with private insurance (n = 153), and Cohort 4 included patients who were uninsured (n = 13). For the second set of analyses, the patients were divided into 2 cohorts: Cohort 1 included patients who were underinsured (Medicaid and uninsured, n = 119) and Cohort 2 included patients who were well-insured (Medicare or private insurance, n = 238). The third set of analyses divided the patients into 2 cohorts: Cohort 1 included patients with private insurance (n = 153) and Cohort 2 included patients with all other types of insurance (Uninsured, Medicare, Medicaid, n = 204). Additionally, Medicare patients were compared directly to privately insured patients. Of the 357 people (222 Female, 135 Male) who met inclusion criteria, there were 13 uninsured, 85 Medicare, 106 Medicaid, and 153 privately insured patients.

Statistical Analysis

After patients were divided into groups based on insurance status, the average pain and standard error of the mean was calculated for each cohort. Means were compared using a t-test. The sex of the patients, along with the percent of patients who were smokers, used opiates prior to the procedure or used antiinflammatories prior to the procedure were compared across cohorts with a chi-square test.

Following the univariate analyses described above, multivariate linear regression analysis was performed to control for the effect of multiple categorical and continuous variables while assessing the contribution of insurance status to pain scores. This model included BMI and age as continuous variables and sex, race, preoperative opioids, and nonsteroidal antiinflammatory drugs, patient smoking status, and surgery procedure type as categorical variables. A 2-sided P-value of ≤ .05 was considered statistically significant.

Results

Demographics and Operative Characteristics

Three-hundred and fifty-seven patients underwent upper extremity surgery during the study time period and met inclusion and exclusion criteria. The frequency of various types of procedures performed differed between insurance groups (P < .001). There was no statistically significant difference in preoperative antiinflammatory use in patients with varying insurance types (P = .811), but there was a difference in preoperative opiate use (P = .012). The percentage of current smokers was significantly higher in uninsured and Medicaid patients compared to other insurance groups (P < .001). However, the percentage of patients who had never smoked was similar between groups. There were no statistically significant sex differences between the Medicaid and Medicare groups; however, the private and uninsured groups had a higher proportion of males when compared to Medicaid (P < .001). Patient demographic and clinical data can be found in Table 1.

Table 1.

Patient Demographics and Clinical Characteristics.

	Payer status				P value
Variable Studied	Medicaid	Medicare	Private	Uninsured
Number	106	85	153	13
Age at surgery in years, mean (SD) (P value) a	47.6 (12.7)	63.8 (12.3) (<.001) a	46.5 (16.0) (.551) a	33.9 (13.1) (.001) a
Sex
Male	27.4%	31.8%	45.1%	76.9%	<.001
Female	72.6%	68.2%	54.9%	23.1%
Pain medications
Antiinflammatories	33.0%	36.9%	30.7%	30.8%	.811
Opiates	50.9%	56.0%	36.0%	38.5%	.012
Smoking status
Never smoker	67.0%	63.9%	74.3%	61.5%	<.001
Former smoker	11.3%	28.9%	17.8%	7.7%
Current smoker	21.7%	7.2%	7.9%	30.8%
Type of procedure
Soft tissue	50.9%	28.2%	44.4%	46.2%	<.001
Fracture/bony repair	34.9%	24.7%	44.4%	46.2%
Joint replacement	4.7%	37.7%	5.9%	7.7%
Nerve repair	3.8%	3.5%	0.7%	0.0%
Arthroscopy	5.7%	5.9%	4.6%	0.0%

Note. SD = standard deviation.

^aP value in comparison to patients with Medicaid insurance.

Please note bold values indicate significance.

Medicaid Versus Private

On univariate analysis, the average preoperative VAS-pain score was higher in the Medicaid group when compared to the private group at all time points, except at the 6-month follow-up appointment (Table 2; Figure 1). Multivariate linear regression confirmed these findings (Table 3), finding a mean adjusted increase in VAS pain scores in Medicaid patients (compared to privately insured patients) of 0.51 preoperatively (95% confidence interval [CI]: 0.26, 0.77), 0.38 at 2 weeks (95% CI: 0.15, 0.62), 0.39 at 1 month (95% CI: 0.11, 0.66), and 0.79 at 3 months (95% CI: 0.49, 1.09).

Table 2.

Univariate Comparison of VAS Pain Scores in Medicaid Versus All Other Modes of Insurance (Medicare, Private, Uninsured).

Payer status	Preoperative, VAS ± SEM (P value) a	Postoperative, VAS ± SEM (P value) a
	(n = 347)	2 weeks (n = 319)	1 month (n = 271)	3 months (n = 178)	6 months (n = 52)
Medicaid	6.9 ± 0.3	4.1 ± 0.3	3.5 ± 0.4	3.9 ± 0.5	3.1 ± 0.9
Medicare	5.8 ± 0.3 (.018)	3.3 ± 0.3 (.060)	2.9 ± 0.3 (.221)	2.4 ± 0.5 (.006)	2.2 ± 0.9 (.407)
Private	5.1 ± 0.3 (<.001)	2.7 ± 0.2 (<.001)	2.1 ± 0.2 (.001)	1.5 ± 0.2 (<.001)	1.5 ± 0.4 (.103)
Uninsured	5.0 ± 0.8 (.044)	2.3 ± 0.7 (.029)	2.8 ± 0.7 (.466)	1.7 ± 0.5 (.015)	1.0 ± 1.0 (.325)

Note. VAS = Visual Analogue Scale; SEM = standard error of the mean.

^aP values in comparison to patients with Medicaid insurance; SEM; significant P values shown in bold.

Figure 1.

This graph displays mean VAS-pain scores for all payer types (Medicaid, Medicare Private, Uninsured) at multiple time points (preoperative, 2 weeks, 1 month, 3 months, 6 months).

Note. The standard error of the mean is displayed in brackets. An asterisk denotes if there was a statistically significant difference between Medicaid and other payer types on univariate analysis. VAS = Visual Analogue Scale.

Table 3.

Multivariate Differences of VAS Pain Scores in Medicaid Versus All Other Modes of Insurance (Medicare, Private, Uninsured).

Payer status	Preoperative VAS a	Postoperative VAS a
		2 weeks	1 month	3 months	6 months
Medicare	0.99 (0.47 to 1.50) (<.001)	0.46 (–0.06 to 0.99) (.086)	0.52 (–0.10 to 1.13) (.098)	0.94 (0.15 to 1.73) (.021)	1.70 (–0.21 to 3.60) (.080)
Private	0.51 (0.26 to 0.77) (<.001)	0.38 (0.15 to 0.62) (.002)	0.39 (0.11 to 0.66) (.007)	0.79 (0.49 to 1.09) (<.001)	0.50 (–0.23 to 1.229) (.169)
Uninsured	0.79 (–1.10 to 2.68) (.409)	1.35 (–0.49 to 3.20) (.149)	0.51 (–1.65 to 2.66) (.640)	1.70 (–0.21 to 3.60) (.080)	–0.240 (–9.36 to 8.83) (.951)

Note. VAS = Visual Analogue Scale.

^aBeta value of multivariate linear regression with 95% confidence intervals in first set of parentheses followed by P-value. Unstandardized beta coefficient represents unit change in the outcome variable if the predictor variable is positive. For example, a significant coefficient of 0.79 for VAS-pain means that, on average, Medicaid is associated with an increase in pain of 0.79 units. Significant P-values shown in bold.

Medicaid Versus Medicare

On univariate analysis, Medicaid patients experienced significant increases in pain compared to Medicare patients at the preoperative clinic visit and at the 3-month clinic visit (Table 2; Figure 1). Multivariate analysis found adjusted differences of 0.99 preoperatively (95% CI: 0.47, 1.50) and 0.94 at 3 months (95% CI: 0.15, 1.73). There were no differences at 2 weeks, 1 month, or at 6 months on univariate or multivariate analysis (Table 2; Figure 1, Table 3).

Medicaid Versus Uninsured

Compared to uninsured patients, Medicaid patients had unadjusted increases in VAS pain scores preoperatively, at 2 weeks and at 1 month (Table 2; Figure 1). However, multivariate regression found no differences in pain scores at any time point in uninsured versus Medicaid patients (Table 3).

Underinsured (Uninsured/Medicaid) Versus Adequately Insured (Private/Medicare)

When assessing VAS-pain scores in underinsured versus adequately insured patients, there were unadjusted VAS-pain score increases preoperatively, at 2 weeks, at 1 month, and at 3 months in underinsured patients (Table 4). Multivariate regression confirmed the differences of univariate analysis, finding adjusted differences of 1.34 preoperatively (95% CI: 0.67, 2.01), 0.89 at 2 weeks (95% CI: 0.26, 1.53), 1.05 at 1 month (95% CI: 0.35, 1.79), and 1.84 at 3 months (95% CI: 0.95, 2.73). There were no differences in VAS-pain scores at 6 months.

Table 4.

Comparison of Pain in Underinsured (Medicaid and Uninsured) Versus Adequately Insured (Medicare and Private) Patients.

	Payer status, mean (SEM)				P value	Beta coefficient a	P value
PAIN VAS	(n)	Underinsured	(n)	Adequately insured
Preoperative VAS	(115)	6.7 (0.3)	(232)	5.3 (0.2)	<.001	1.34 (0.67 to 2.01)	<.001
Postoperative VAS
2 weeks	(107)	3.9 (0.3)	(212)	2.9 (0.2)	.003	0.89 (0.26 to 1.53)	.006
1 month	(85)	3.4 (0.4)	(186)	2.4 (0.2)	.007	1.05 (0.35 to 1.79)	.004
3 months	(49)	3.4 (0.4)	(129)	1.8 (0.2)	.001	1.84 (0.95 to 2.73)	<.001
6 months	(17)	2.9 (0.8)	(35)	1.8 (0.4)	.130	1.31 (–0.62 to 0.25)	.178

Note. SEM = standard error of the mean; VAS = Visual Analogue Scale.

^aMultivariate Linear Regression unstandardized beta coefficient with 95% confidence intervals, defines the adjusted difference of pain for underinsured patients in comparison to adequately insured patients when taking in consideration confounding factors included in model.

Significant p values are shown in bold.

Private Versus Uninsured/Medicaid/Medicare

Privately insured patients experienced decreased pain scores compared to the other cohorts at all time points except the 6-month follow-up appointment. Multivariate regression confirmed univariate analysis, finding adjusted differences of 0.81 preoperatively (95% CI: 0.16, 1.45), 0.71 at 2 weeks (95% CI: 0.01, 1.34), 0.75 at 1 month (95% CI: 0.47, 2.07), and 1.27 at 3 months (95% CI: 0.95, 2.73).

Private Versus Medicare

On univariate analysis, there were no differences in privately insured patients when compared to Medicare patients, although there was a trend toward significance at some time points (P values .051-.084). However, multivariate analysis did not find a trend toward significance and confirmed there were no statistical differences among these groups.

Pain Improvement

There was no statistically significant difference in pain improvement found between any groups when comparing any of the cohorts above (Table 5). These results were confirmed with multivariate linear regression (all P-values > .05).

Table 5.

Univariate Comparison of Improvement in VAS Pain Scores in Medicaid Versus All Other Modes of Insurance (Medicare, Private, Uninsured) from Baseline.

Payer status	ΔVAS ± SEM (P value) a
	Baseline to 2 weeks (n = 319)	Baseline to 1 month (n = 271)	Baseline to 3 months (n = 178)	Baseline to 6 months (n = 52)
Medicaid	2.7 ± 0.4	3.2 ± 0.5	3.5 ± 0.5	4.1 ± 1.0
Medicare	2.4 ± 0.4 (.660)	2.8 + 0.5 (.540)	3.3 + 0.5 (.745)	3.2 ± 1.2 (.534)
Private	2.1 ± 0.3 (.214)	2.8 ± 0.4 (.515)	3.5 ± 0.4 (.990)	3.7 + 0.8 (.724)
Uninsured	3.3 ± 0.8 (.585)	2.1 ± 1.1 (.372)	3.5 ± 1.1 (.975)	2.5 ± 0.5 (.571)

Note. VAS = Visual Analogue Scale; SEM = standard error of the mean.

^aP values in comparison to patients with Medicaid insurance; SEM.

Discussion

As the number of patients receiving insurance through Medicaid increases, it is important to further understand how patient insurance status relates to outcomes following surgery. Numerous studies have shown a correlation between insurance status and increased mortality and complication rates across multiple surgical fields.^10-16 An often-overlooked outcome in these studies has been postsurgical pain relief. The purpose of this study was to investigate the relationship between insurance status and patient reported pain both before and after upper extremity surgical procedures.

Our results showed that Medicaid patients reported higher levels of preoperative and early postoperative pain relative to patients with Medicare or private insurance; however, we found no significant difference in reported pain levels between patients in Medicaid and non-Medicaid groups at 6 months postoperatively. These findings can be used to help guide patient expectations regarding pain improvement in the early and intermediate postoperative periods. Additionally, as healthcare systems are becoming increasingly dependent on patient-reported outcomes, including pain, it is important to consider that differences may exist in subjective pain depending on insurance status.

While we found that Medicaid patients reported higher levels of preoperative and early postoperative pain, the exact reasons for this finding are unclear. Prior research has suggested that individuals of a lower SES experience higher levels of pain. Recent research has suggested that patient perceptions can largely affect functional recovery and outcomes after orthopedic injuries. ⁵ A study by Zelle et al, ⁵ found ethnic differences in patient perceptions of isolated orthopedic injuries, and further suggested that these differences in perceptions may affect perceptions of pain and recovery. When considering the racial and ethnic differences between patients who are insured versus underinsured, it stands to reason that differences in perception of care may play a role in patient-reported pain scores both pre- and postoperatively.

Interestingly, we found that both Medicare and Medicaid patients reported higher levels of opiate use relative to patients with private insurance and those with no insurance. A number of studies have demonstrated that preoperative opiate usage is associated with increased postoperative pain, which may help to explain why Medicaid patients report higher levels of pain in the early postoperative period, although preoperative opioid use was included in our multivariate regression.^18-21 Additionally, it is important to consider that underinsured patients often have decreased access to orthopedic care and have poorer continuity of care relative to patients with private insurance.^1-4 It is possible that Medicaid patients present with more severe conditions due to these barriers to orthopedic care, including increased difficulty scheduling appointments and delays in receiving diagnoses. Future studies investigating the relationship between insurance status and the severity and duration of patient symptoms could provide further insight into why Medicaid patients experience higher levels of preoperative pain.

A recent study investigated the effect of Medicaid payer status on function following repair of massive rotator cuff tears. ¹⁵ Results of that study demonstrated a significant difference in preoperative pain and functioning with Medicaid patients reporting more pain and decreased function relative to non-Medicaid patients (ASES American Shoulder and Elbow Surgery Score, PENN Penn Shoulder Score, SVV Subjective Shoulder Score), but no significant differences in these outcomes postoperatively. Similar to these results, our study demonstrates no significant differences at 6 months postoperatively. The results of these studies suggest that although underinsured patients may report increased pain and functional limitations in the pre- and early postoperative periods, these measures equilibrate after the intermediate postoperative period.

This study has a number of strengths as well as several limitations. We included a large sample size with a broad inclusion criteria in an attempt to increase the generalizability of our results to the majority of patients who undergo upper extremity orthopedic surgery. Additionally, we used the VAS as a measure of subjective pain at each time interval, which is a measure that has been validated as a reliable measure of subjective pain. ²² Despite these strengths, our results should be interpreted with caution. As this was a retrospective design, the results are subject to potential biases associated with the nature of the study. We included a broad range of upper extremity procedures, and found that there were differences in the types of procedures performed on each cohort. It is therefore possible that differences in pain outcomes may be partly attributable to differences in either the orthopedic condition or procedure performed; however, we performed multivariate analysis to help control for this confounding. In addition, outcomes may be compounded by the duration of symptoms prior to treatment, which was not accounted for in the current study; a longer time to treatment for underinsured patients could affect pain levels and therefore manifest as increased preoperative and postoperative pain scores. Further information about patient function would provide insight into whether the increased pain experienced by Medicaid patients was linked to more debilitating injuries. Another notable limitation of the study is the attrition rate seen across all groups. While greater than 70% of patients in all insurance groups completed VAS scores up to 1 month postoperatively, this percentage decreased to a mean of 56.7% at 3 months and a mean of 15.1% completion rate at 6 months postoperatively. This is because certain types of surgery, like carpal tunnel release, do not require follow-up at 6 months. These loss-to-follow-up rates could introduce a source of selection bias that may have influenced pain scores, although rates were similar across insurance types, minimizing the possibility of this limitation affecting intergroup associations. Finally, although we were able to obtain diverse and well-matched patient groups for this study, it was limited to those patients seeking treatment from a single surgeon and may not be applicable to all practice situations. We acknowledge that a number of factors likely influence pain scores and do not attempt to attribute these differences in outcomes to patient insurance status. Rather, we feel that it is important to acknowledge that differences in pain outcomes exist and are associated with insurance status. These results are important to consider when managing patient expectations.

In conclusion, this study demonstrates that although Medicaid patients reported higher levels of preoperative and early postoperative pain up to 3 months postoperatively compared to patients with private insurance, all patients showed significant improvement in pain, and no differences in pain level were found beyond the 6-month postoperative period.