Changes in patient visits following the implementation of insurance billing at a sexually transmitted diseases clinic in a Medicaid Expansion state

Abstract

Background:

Medicaid expansion has led to unique opportunities for sexually transmitted disease (STD) clinics to improve the sustainability of services by billing insurance. We evaluated changes in patient visits after the implementation of insurance billing at a STD clinic in a Medicaid expansion state.

Methods:

The Rhode Island STD Clinic offered HIV/STD screening services at no cost to patients until October 2016, when insurance billing was implemented. Care for uninsured patients was still provided for free. We compared the clinic visits in the pre-insurance period to the post-insurance period using t-tests, Poisson regressions, and a logistic regression.

Results:

A total of 5,560 patients were seen during the pre- (N=2,555) and post-insurance (N=3,005) periods. Compared to the pre-insurance period, the post-insurance period had a significantly higher average number of patient visits/month (212.9 vs. 250.4, p=0.0016), including among patients who were Black (36.8 vs. 50.3, p=0.0029), Hispanic/Latino (50.8 vs. 65.8, p=0.0018), and insured (106.3 vs. 130.1, p=0.0025). The growth rate of uninsured (+0.10 vs. +4.11, p=0.0026) and new patients (−4.28 vs. +1.07, p=0.0007) also increased between the two time periods. New patients whose first visit was before the billing change had greater odds (adjusted odds ratio: 2.68, 95% CI = 2.09–3.44, p-value<0.0001) of returning compared to new patients whose first visit was after the billing change.

Conclusions:

Implementation of insurance billing at a publicly-funded STD clinic, with free services provided to uninsured individuals, was associated with a modest increase in patient visits and a decline in patients returning for second visits.

Brief Summary:

A study of patient volume at a publicly-funded STD clinic found that implementation of insurance billing may provide a sustainable model of STD care with minimal changes on patient volume.

Introduction

Sexually transmitted diseases (STD) including syphilis, gonorrhea, and chlamydia continue to be a major public health burden in the United States (US). There are an estimated 20 million incident STD cases in the US each year, which costs nearly $16 billion annually to treat.¹ Publicly-funded STD clinics have typically offered services at no/low cost to patients. However, the sustainability of these services has been limited by insufficient public health funding.^2–4 Medicaid expansion in many states has offered an opportunity to implement insurance billing practices that could provide a viable financial model for these clinics.^4,5 However, previous studies examining the impact of implementing insurance billing on STD service utilization, including models that have increased out-of-pocket costs, charged a flat fee, or required mandatory insurance billing, have reported significant reductions in patient visits, especially among high-risk populations.^6–8

The Rhode Island STD clinic is the only publicly-funded STD clinic in the state and provides safety-net testing and treatment services. Historically, STD testing and treatment services were provided to patients for free. With the expansion of Medicaid and its patient enrollment, over 95% of Rhode Island residents had insurance as of 2015.⁹ To improve financial sustainability, the Rhode Island STD Clinic began billing people who had insurance in October 2016, while still providing free care to people without insurance. The objective of this study was to examine changes in patient visits after the implementation of insurance billing at a STD clinic in a Medicaid expansion state. Specifically, we compared the pre- and post-billing time periods’ average monthly patient visits and STD screenings as well as the average monthly growth rate of patient visits and STD screenings. We also modeled a logistic regression to examine if patients were as willing to return for a second visit within 6 months of their first visit following the implementation of the new billing system.

Methods

We first reviewed the demographic characteristics, behaviors, insurance status (yes/no), insurance type, and use of insurance for all first patient visits from October 2015 to October 2017. We compared the 12 months prior to insurance billing (October 2015-September 2016) relative to the 12 months after (November 2016-October 2017). The insurance status of all patients was based on self-report and also validated using our medical record system. We excluded the one-month during which insurance billing was implemented (October 2016) to account for a transition period.⁶ We compared the rate of patient visits and STD screenings per month during the pre- and post-insurance periods using t-tests. The number of STD clinic visits was increasing over time beginning in the pre-insurance period. Therefore, we then compared the average growth rate of patient visits and STD screenings per month in the two periods using Poisson regression models, which included the period (pre- or post-insurance), the month, and an interaction term for the two variables. This approach was used to account for the increase in patient visits per month by measuring the change in rate instead of the number of patient visits. We also evaluated each comparison among new patients and specific demographic subgroups by gender, race, ethnicity, men who have sex with men (MSM), and insurance status. The Bonferroni correction was used to adjust the critical p-value level of significance to 0.003 (0.05/17). To address potential misclassification of health insurance status, we conducted a sensitivity analysis in which the insurance status of 200 patients who self-reported not having health insurance were changed to having insurance.

In a secondary analysis, we evaluated whether new patients had the same odds of having a repeat visit after the billing change. We compared the odds of patients returning for a second visit among patients whose first visits were between November 2015 and April 2016 relative to patients whose first visits were between November 2016 and April 2017. We used a logistic regression model (to avoid underestimating standard errors with a maximum likelihood model) with patient fixed effects and half-year fixed effects.¹⁰ Patient fixed effects controlled for time-invariant patient characteristics, such as gender, ethnicity, and race.

All analyses were conducted in SAS 9.4 (Cary, North Carolina). The local institutional review board approved study protocols.

Results

A total of 5,560 patients were seen during the pre- (N=2,555) and post- (N=3,005) billing time periods. The total monthly patient visits from October 2015 to October 2017 are illustrated in Figure 1. The Clinic saw an average of 212.9 patients/month with an average growth rate of +1.84 patients/month during the pre-insurance period, and saw an average of 250.4 patients/month with an average growth rate of +4.02 patients/month during the post-insurance period (Table 1). There was a statistically significant increase in the average number of patient visits per month (212.9 vs. 250.4, p=0.0016) but no difference in the growth rate of the clinic between the two periods (+1.84 vs. +4.02 patients/month, p=0.34). There was no difference between the average number of new patients per month (172.0 vs. 168.1, p=0.63), however, the growth rate of all patient visits increased from −4.28 to 1.07 patients/month (p=0.0007).

Figure 1.

Total Number of Patient Visits by Month at the Rhode Island STI Clinic, October 2015 – October 2017.

Note: Visits that occurred in October 2016 were excluded from any analysis because the billing change occurred during that month.

Table 1.

Changes in average monthly patient and STI screening visit and growth rates before (October 2015-September 2016) and after (November 2016-October 2017) the implementation of insurance billing at the Rhode Island STI Clinic.

	Total Visits (n=5,560)		Average monthly rate			Average monthly growth rate
	Time period 1	Time period 2	Time period 1	Time period 2	P-value	Time period 1	Time period 2	P-value
Total Number of Visits	2,555	3,005	212.9	250.4	0.0016	1.84	4.02	0.3396
New Patients	2,064	2,017	172.0	168.1	0.6284	−4.28	1.07	0.0007
Gender
Female	653	750	54.4	62.5	0.0392	0.15	0.94	0.4343
Male	1,890	2,190	157.5	182.5	0.0094	1.63	3.02	0.4928
Transmen	0	3	0	0.3	0.1911		0.75
Transwomen	11	5	0.9	0.4	0.1875	−0.03	0.00	0.9681
MSM	892	1,044	74.3	87.0	0.0290	0.81	2.32	0.2366
Race
White	1,273	1,422	106.1	118.5	0.0614	0.00	1.83	0.1723
Black	441	603	36.8	50.3	0.0029	0.64	2.19	0.1545
Other Race	841	980	70.1	81.7	0.0267	1.19	0.05	0.2252
Hispanic or Latino	610	789	50.8	65.8	0.0018	0.68	0.38	0.6226
Insurance Status
Insured	1,276	1,561	106.3	130.1	0.0025	1.74	0.14	0.1587
Not Insured	1,278	1,398	106.5	116.5	0.2119	0.10	4.11	0.0026
STI Screenings
HIV	2,272	2,408	189.3	200.7	0.2063	2.13	1.85	0.8330
Hepatitis C	1,137	1,010	94.8	84.2	0.0652	0.65	1.15	0.5866
Gonorrhea	2,421	2,836	201.8	236.3	0.0007	1.67	2.64	0.7130
Chlamydia	2,419	2,786	201.6	232.2	0.0036	1.55	3.88	0.2613
Syphilis	2,309	2,601	192.4	216.8	0.0160	1.91	2.87	0.6930

Compared to the pre-insurance period, the post-insurance period had a significantly higher average number of patients who were Black (36.8 vs. 50.3, p=0.0029), Hispanic/Latino (50.8 vs. 65.8, p=0.0018), and insured (106.3 vs. 130.1, p=0.0025), and the average number of gonorrhea screenings increased from 201.8 to 236.3 (p=0.0007). The clinic also saw an increase in the average growth rate among uninsured patients (+0.10 vs. +4.11, p=0.0026). No differences were found by gender, MSM status, or race. There were no differences in the growth rate of screenings for HIV or other STDs between the two periods.

The results of the sensitivity analysis were similar to that of the primary analysis (Table 2). There was an increase in the average number of visits made by insured patients (106.3 vs. 130.1, p=0.0025) and an increase in the average growth rate among uninsured patients (−0.57 vs. +4.29, p=0.0001).

Table 2.

Sensitivity Analysis - Changes in average monthly patient visit and growth rates among insured and insured patients before (October 2015-September 2016) and after (November 2016-October 2017) the implementation of insurance billing at the Rhode Island STI Clinic.

	Total Visits (n=5,560)		Average monthly rate			Average monthly growth rate
	Time period 1	Time period 2	Time period 1	Time period 2	P-value	Time period 1	Time period 2	P-value
Insurance Status
Insured	1430	1695	119.2	141.3	0.0054	2.39	0.00	0.0512
Not Insured	1124	1264	93.7	105.3	0.1319	−0.57	4.29	0.0001

New patients whose first visit was before the billing change (November 2015-April 2016) had greater odds (odds ratio: 2.68, 95% CI = 2.09–3.44, p-value<0.0001) of returning compared to new patients whose first visit was after the billing change (November 2016-April 2017) (Figure 2).

Figure 2.

The first and second visits of new patients who visited the Rhode Island STI Clinic before (November 2015-April 2016) and after (November 2016-April 2017) the implementation of insurance billing.

Discussion

This study sought to evaluate STD clinic outcomes after insurance billing implementation in a Medicaid expansion state. Importantly, patients without insurance were still provided care for free. Insured patients who sought care at the clinic were encouraged to use their insurance but were given the option to pay out-of-pocket if they preferred not to use their insurance, which was cost prohibitive in almost all instances. There are STD clinics in the US that give insured patients the option to access free services or pay a flat-fee without billing their insurance¹¹; however, several clinics using these models have reported decreased patient volume.^6–8 More research is urgently needed to understand how to reduce barriers of care, such as concerns with confidentiality and increased out-of-pocket costs, while also maintaining the clinics’ financial sustainability.

We found that there was a significant increase in the average number of patient visits in the post-insurance period compared to the pre-insurance period. These findings demonstrate that implementation of insurance billing at a STD clinic did not lead to reductions in patient visits or STD screenings, and may be a sustainable model of providing STD testing and treatment services to underserved populations.

While the Affordable Care Act (ACA) has increased insurance coverage across the US, there is still continued demand for STD clinics.¹² Several studies have found that patients continue to seek services at STD clinics, even among insured patients who can access health care elsewhere.^12,13 These studies reported that patients preferred STD clinics because of their convenience and expertise in sexual health. However, previous studies have demonstrated that charging a flat fee may deter both insured and uninsured patients from accessing STD services.^7,14,15 Many insured patients are unable to afford copayments or deductibles for STD services.¹⁴ Additionally, a significant subset of people, many from underserved populations (e.g., racial/ethnic minorities), will likely remain uninsured despite ACA implementation and Medicaid expansion.¹⁶ These limitations of the healthcare system emphasize the need for sustainable payment structures to ensure that STD services are accessible by both insured and uninsured patients.¹⁵

Differential health care access between white communities and communities of color has been repeatedly demonstrated and is a significant driver of health disparities.^17–19 Previous studies have found that ethnic and racial minorities are at increased risk for STDs, are less likely to have established sources of primary care,^12,20,21 and that the rate of uninsured visits to community health centers decreased among all racial and ethnic groups following the expansion of Medicaid.²² In other STD clinic settings, institution of flat fees⁶ resulted in an immediate drop in patient visits and STD diagnoses. This suggests that even modest out-of-pocket expense may differentially affect economically disadvantaged populations. An important difference in the approach at the Rhode Island STD Clinic is that uninsured patients were still provided care for free. A financial model in which insured patients are billed for services, and uninsured patients are covered for free, may be a sustainable and effective model of providing STD services.

We found that new patients whose first visit was before the billing change had greater odds of returning to the clinic compared to new patients whose first visit was after the billing change. Patients whose first visit was after the implementation of the billing system may have been discouraged from returning to the clinic after being informed of the new policy. Studies have found that patients may be hesitant to use their insurance at STD clinics because of privacy concerns or increased out-of-pocket costs,¹⁴ which may provide insight into why insured patients were less likely to return if they were notified of the new policy at their first visit. Surprisingly, the decrease was more significant among uninsured patients, even though uninsured patients were not required to pay out-of-pocket costs. Future work is needed to determine barriers to making return visits in this patient population.

A limitation of this study is that the analysis compares two finite time periods and did not specifically analyze the causal relationship between the implementation of the billing change and patient volume. Despite this, the absolute number of visits increased in the post-insurance period, as did the monthly visit rate of uninsured patients. Our findings suggest that billing patients with insurance and providing uninsured patients care at no cost may diminish the financial barriers that patients perceive when seeking STD treatment and prevention services. Another limitation is how we determined patients’ health insurance status. All patients self-reported their insurance status, which was verified using our medical record system. However, there is a chance that an insured patient who has never engaged with our healthcare system may have reported being uninsured. This would result in a potential overestimation of uninsured individuals. To address this limitation, we conducted a sensitivity analysis in which we randomly selected 200 patients who self-reported being uninsured and we switched their insurance status (Table 2). The results of the sensitivity analysis remained consistent with the results of our main analysis. Finally, we also acknowledge that there may be other factors unrelated to the billing change that could have affected patient volume, such as seasonal changes in patient volume. In order to address this limitation, we compared the entire calendar year before the billing change to the entire year following, which allowed us take seasonal changes into consideration.

In conclusion, STD clinics provide important safety-net services to underserved populations. Implementation of insurance billing may provide a sustainable model of STD care with minimal impact on patient volume. Mechanisms to ensure access to care for uninsured and unserved populations are critical for effective delivery of STD services.