Cost and Utilization of Retail Clinics vs. Other Providers for Treatment of Pediatric Acute Otitis Media

Abstract

A common acute condition seen by providers in retails clinics is the evaluation and treatment of acute otitis media (AOM) in children younger than age 20. Annual direct treatment costs for AOM were US $5.3 billion in 1998 dollars. Based on the experience of a large retail pharmacy employer, the authors compared AOM episodes in covered dependents younger than age 20 in retail clinic states to those in states without retail clinic access. Relative costs as well as frequency of visits and antibiotic prescriptions were analyzed for both retail clinic-based, and non–retail clinic-based episodes. Rates of AOM episodes were lower in retail clinic than in non–retail clinic states (62.5 vs. 76.9 per 1000 members per year; P < .0001). The average number of visits per episode was similar between retail clinic and non–retail clinic states (1.417 vs. 1.430, respectively; P = 0.657), suggesting that retail clinics do not result in an increase in overall utilization. On a risk-adjusted basis, retail clinic episodes cost approximately $30–$130 less than community episodes, depending on year. In retail clinic states, the antibiotic prescription fill rate was 95.4% for retail clinic episodes and 82.8% for community episodes, consistent with rates in the literature. This study confirms results of earlier studies that retail clinics are a less costly setting than the community for the treatment of episodes of otitis media There also is little evidence that retail clinics lead to duplication of services (patients receiving follow-up care in other settings).

Background

Retail clinics are an increasingly popular source of care for acute episodic needs, such as acute ear infections. Located in pharmacies, grocery stores, and other retail locations, the attraction of retail clinics is their convenience—the expanded hours of operation and the ability to access care without a wait or even, at times, an appointment. Mehrotra et al¹ reported that there were nearly 1000 retail clinics in 2009. According to a report by the consulting firm Accenture,² this number had grown to 1400 by 2012 and is estimated to reach 3000 by 2016.³

The range of services offered by retail clinics is more limited than physician offices, emergency departments (EDs), or urgent care centers, generally consisting of preventive services such as immunizations, health screenings and physicals, treatment of upper respiratory tract infections, and treatment of common illnesses and minor injuries (although recently some clinic operators have begun to offer care for patients with chronic diseases). In addition, some retail clinics set a minimum age such as 18 months for the patients whom they will treat. Some physician organizations have raised concerns about the quality of care delivered by retail clinics. In addition, payers are concerned that the care given in retail clinics is not substitutive but duplicative: patients, having first used the service of a nurse practitioner in the retail setting may subsequently seek care from a physician in a physician practice, ED, or urgent care center. Although the literature on quality of care in the retail setting has been growing, the impact on cost and utilization has not been subject to the same analysis. A recent study by Sussman et al⁴ mentions only a single study, that by Rohrer et al.⁵ Although there are a few more published studies, the number tends to be limited. Studies of comparative costs and utilization by patients in retail clinics vs. other providers also suffer from the drawback that they compare (self-selected) patients of retail clinics with those who seek care in other settings. Although this selection bias can be addressed by appropriate statistical methods, the potential for criticism remains (see, for example, correspondence by R. Kochar⁶ commenting on the study by Mehrotra et al).¹ The motivation for the present study includes the ability to compare the experience of patients in states with access to retail clinics with that of a natural comparison population drawn from states without access to retail clinics.

Otitis Media

Acute otitis media (AOM) is gaining increasing public health significance amid the growing US health crisis involving microbial resistance to antibiotics.^7
–9 It is the second most common disease of childhood after upper respiratory infection, with approximately 80% of AOM patients being children younger than age 15; more than 80% of children develop acute episodes by age 3^10,11 and about 40% suffer 6 or more acute episodes by age 7.¹² In the 1990s, annual direct treatment costs for AOM were estimated to be between US $4.0 and $5.0 billion in 1998 dollars.^13,14 More recent estimates are lower but still significant.¹⁵

Whereas AOM is an important cause of disability and mortality in developing regions owing to intercurrent malnutrition, infectious disease, and overcrowding, in industrialized countries AOM has high rates of spontaneous resolution¹⁶ and low rates of serious acute medical complications,^17,18 even when left untreated.¹² Nevertheless, the favorable natural history of AOM is at striking odds with its acute clinical presentation, especially among infants and younger children. Typical symptoms include high fever, severe pain, and “irritability.” Despite emerging evidence regarding the low relative individual risk associated with AOM and high population risk associated with antibiotic resistance, AOM remains one of the most treatment-intensive and costly pediatric disorders in the United States, as well as being the third leading cause of pediatric ED visits. AOM has also been the leading cause of outpatient pediatric antibiotic prescriptions for 2 decades.^18

–21

Previous Studies of Cost, Utilization, and Quality

Several studies have compared the cost of retail clinic services with those of other settings (Mehrotra et al,¹ Rohrer et al,⁵ Thygeson et al,²² and Sussman et al).⁴ These results (as well as the methods used for comparison) are summarized in online Table 1S, available at www.liebertpub.com/pop.

Objectives

This study had 3 objectives:

1. Analyze the relative cost of AOM episodes in retail clinic and community settings;

2. Determine the frequency of both episodes and follow-up community and emergency services in episodes that begin in retail clinics (the substitution vs. duplication hypothesis); and

3. Evaluate the frequency of antibiotic prescription fills in retail clinic and community settings.

This study examined an employee population with access to physician office, urgent care, ED, outpatient hospital, and retail clinic services. Because of state laws governing the practice of medicine, not all employees had access to a retail clinic. Thus, unlike prior studies, the present study population was divided into 3 cohorts: those members who have access to retail clinics (members in Florida, Illinois, Texas, and Wisconsin) and use them; those members who have access to a retail clinic and choose to use an alternative setting; and those members who live in states without retail clinics (California, Michigan, Minnesota, Iowa, Oklahoma, and Nebraska) who are unable to access a clinic. All members were younger than age 20. Although the members in states with access to a retail clinic may experience some self-selection, the comparison group (those who have no access) are not affected by self-selection with respect to clinic services. Similar to earlier studies, the study team examined the cost of an episode of care anchored by an AOM diagnosis. However, patients are classified according to a larger number of categories. Thus, episodes that begin in a retail clinic, for example, but that later involve another setting are classified as “mixed.”

Data

This study examined the utilization and cost of episodes of AOM among members younger than age 20 of a large national pharmacy company with an extensive retail clinic network, using administrative claims data only. Incurred claims data for January 2009 through September 2012 (including runout through November 2012) were obtained from the company's data aggregator. The study team obtained eligibility information at the member level that included age band, sex, start and end dates for eligibility, 3-digit home zip code, relationship to employee (employee, spouse, or dependent), insurance carrier, and plan. (For privacy reasons ages were banded, implying that the study team did not have specific ages of employees or dependents.) Because the data set was derived from administrative claims, information about race and household income of the members was not available. Eligible lives by year and state grouping are shown in Table 1. The distribution of populations between retail clinic states and non–retail clinic states are similar in terms of sex; the non–retail clinic states have relatively more members younger than age 10.

Table 1.

Total Eligible Lives by Year and Clinic Access

Calendar Year	Retail Clinic States	Non–Retail Clinic States	P
2009^*	8,590	2881
% Male	50.9%	50.6%	P > .05
% Less than 10^***	44.4%	51.0%	P < .0001
2010	21,326	8324
% Male	51.3%	50.9%	P > .05
% Less than 10	48.3%	52.6%	P < .0001
2011	21,269	8428
% Male	51.0%	51.1%	P > .05
% Less than 10	50.7%	55.2%	P < .0001
2012^**	20,959	8427
% Male	51.0%	50.6%	P > .05
% Less than 10	52.0%	56.7%	P < .0001

2009 includes only eligible lives between March–December.

2012 includes only eligible lives between January–June.

***

Specific member age was unavailable. Age data were provided for 2 bands (younger than10; 10 to younger than 20).

Full claims data (all medical and pharmacy claims) were obtained for members with at least 1 medical claim with an International Classification of Diseases, Ninth Revision (ICD-9) diagnosis code of 382.9 (AOM) in the employer's retail clinic (identified by National Provider Identifier or tax ID) or as an office visit (identified by place of service = 11).

Benefit design

In 2009 the employer plan was primarily health maintenance organization based, accounting for the relatively low clinic eligibility in this year. In 2010 (and subsequently) the employer plan consisted primarily of preferred provider organization plans with a $5 co-pay as an incentive to use the employer's (and other) retail clinics for non-preventive services (preventive services were covered at 100% for in-network settings). Community provider services required a deductible and a 20% coinsurance payment.

Methods

The study team performed a retrospective cohort study using medical and pharmaceutical claims incurred between 2009 and 2012. This was a retrospective study of a limited administrative claims data set only. In addition, the data provided in several fields (age, for example) were banded for additional privacy. There was no intervention with human subjects and an Institutional Review Board was not involved. Study cohorts were limited to members younger than age of 20 because of the incidence of AOM in younger patients. The test group consisted of patients in states where the employer's retail clinics were available (Florida, Illinois, Texas, and Wisconsin). In these states the team distinguishes between episodes of AOM that begin in the retail clinic and those that begin elsewhere (the “community”). Although a small number of episodes in either case “migrate” between their initial setting to an alternative setting, the majority of episodes begin and end within the same setting. The control group consisted of patients in states without access to retail clinics (California, Michigan, Minnesota, Iowa, Oklahoma, and Nebraska), either because state laws made the establishment of retail clinics impossible or because the employer had decided, for business reasons, not to establish clinics in these states.

Creation of episodes

The study compares costs and utilization of services during episodes of AOM treated in different settings. An “episode” is defined as the aggregate of all services for the diagnosis, beginning with the first observation of such services (after a “clean” period to distinguish from prior episodes) to the last observation of such services. Included in an episode are “visits” (encounters with a medical professional in either the primary care, ED, outpatient, or retail setting), inpatient admissions, and prescription drug fills. Earlier studies of AOM used the Symmetry-grouper (a proprietary commercial algorithm) to create episodes. The study team did not have access to this algorithm and instead created an episode algorithm as follows: the index event for a defined episode was the first medical claim for a primary diagnosis of unspecified AOM (ICD-9 code 382.9; including inpatient admissions, outpatient services, ED, retail clinic, and physician office claims), following a “clean” period of at least 60 days without any AOM-related medical claim. All services with an AOM-related primary diagnosis (irrespective of the code) were then aggregated into the episode until the patient experienced a “clean” period (no further AOM-related medical claims) of at least 60 days. Pharmacy claims for certain antibiotic drugs (see list in online Appendix 1, available at www.liebertpub.com/pop) also were included in the episode costs, if the fill date was within 3 days of an AOM-related visit. Episodes were created for claims incurred between March 1, 2009, and June 30, 2012, to allow for a minimum of 60 days “clean” period before or after an initial AOM claim.

Risk adjustment

Risk adjustment is a relatively new, widely used technique for normalizing different populations on the basis of their diagnostic, age, and sex-based risk for comparative purposes. (The most recent and widest use of the technique by the Centers for Medicare & Medicaid Sevices is in normalization of insurer populations for reimbursement purposes in Medicare Advantage, Medicare Part D, and the Affordable Care Act exchanges.) Each patient is assigned a “score” based on the patient's age, sex, and diagnoses. An average score is calculated for each population, allowing measures such as cost per member per month for the population to be adjusted based on relative scores. For budgetary reasons, this study used the public domain risk adjuster the Chronic Disability Payment System (CDPS; described in online Appendix 2, available at www.liebertpub.com/pop) from the University of California San Diego.²³ The CDPS grouper is a widely-used, publicly-available algorithm that has been evaluated by the Society of Actuaries (SOA; most recently in Winkleman and Mehmud²⁴). The SOA study showed that the CDPS model does not significantly under-perform other models on 2 measures (R² and MAPE [mean absolute percent error]). It should be noted that the child CDPS model was used for this study, a model that did not form part of the SOA evaluation, but one that would be expected a priori to perform better in the present application than other models constructed for adult risk adjustment.

The observed higher number of visits per episode in the community setting, as well as prior studies, suggests that community provider users may be a higher risk population than retail clinic patients. The study team therefore calculated an average condition-based risk score for each AOM patient using the CDPS grouper. The calculation of the CDPS risk score is described in online Appendix 2, and takes into account a number of factors including demographics (such as age band) as well as all diagnoses present in a patient's claims record (not solely those pertaining to AOM). Risk scores are highly correlated with the severity of patient diagnoses and allow for adjustment for the cost implication of relative severity of different diagnoses.²⁵ The relative risk scores of community and clinic patients were applied to the cost of clinic and community episodes for each population to calculate an adjusted cost.

Results

Statistical analysis

Clinic Utilization

Table 2a shows the number of episodes of AOM by year originating in retail clinics and the community, and those transferring between settings. Table 2b shows similar data, but on a rate per 1000 members per year basis.

Table 2.

Acute Otitis Media (AOM) Episodes

Table 2a: Total AOM Episodes by Year
	Calendar Year	Clinic only	Community only	Clinic -> Community	Community -> Clinic	Total
Clinic States	2009	16	444	0	1	461
	2010	107	1065	12	6	1190
	2011	262	929	20	3	1214
	2012	146	441	4	2	593
	Total	531	2879	36	12	3458
Non–Clinic States	2009	0	203	0	0	203
	2010	1	566	0	0	567
	2011	1	544	0	0	545
	2012	1	327	0	0	328
	Total	3	1640	0	0	1643

Table 2b. Total AOM Episodes per 1000 by Year
Clinic States	2009	2.5	68.9	0.0	0.2	71.6
	2010	5.5	54.6	0.6	0.3	61.1
	2011	13.5	47.8	1.0	0.2	62.5
	2012	14.6	44.1	0.4	0.2	59.3
	Total	9.6	52.0	0.7	0.2	62.5^*
Non–Clinic States	2009	0.0	96.8	0.0	0.0	96.8
	2010	0.1	74.6	0.0	0.0	74.7
	2011	0.1	70.9	0.0	0.0	71.0
	2012	0.2	81.7	0.0	0.0	81.9
	Total	0.1	76.8	0.0	0.0	76.9^*

P < .0001.

Note: Available data do not allow determination of the source of clinic claims in non–clinic states. Because employees were classified according to residence, some employees could live in one state but work in another (clinic) state, or be visiting another state.

Data for 2009 includes claims incurred from March–December only.

Data for 2012 includes claims incurred from January–June only.

Table 2b clearly illustrates the penetration made by the retail clinics over time: although the total number of AOM episodes per 1000 in retail clinic states decreased slightly between 2009 and 2012, the rate per 1000 for retail clinic-based episodes increased from 2.5 to 14.6 per 1000 over the same time period.

The majority of episodes consist of a single visit; the mean number of visits per episode is 1.417 (retail clinic states) and 1.430 (non–retail clinic states). The P value between these 2 means is 0.657.

Most clinic episodes were completed with a single visit; community episodes were more likely to include repeat visits. The distribution of visits per episode is shown in Table 3. The percentages of single-visit episodes are consistent with those observed in the study by Angstman et al.²⁶

Table 3.

Number of Visits per Episode

	Clinic States				Non–Clinic States
	Retail Clinic Episodes		Community Episodes		Community Episodes
Count of Visits	#	Cumulative Distribution	#	Cumulative Distribution	#	Cumulative Distribution
1	491	86.6%	2105	72.8%	1245	75.8%
2	52	95.8%	503	90.2%	249	90.9%
3	21	99.5%	160	95.7%	78	95.7%
4	1	99.6%	54	97.6%	32	97.6%
5	1	99.8%	33	98.8%	17	98.7%
6	1	100.0%	18	99.4%	5	99.0%
7		100.0%	7	99.6%	9	99.5%
8	-	100.0%	7	99.9%	6	99.9%
9	-	100.0%	1	99.9%	1	99.9%
10+	-	100.0%	3	100.0%	1	100.0%
Total	567		2891		1643
Mean Visits/ Episode	1.187^*		1.462^*		1.430^*

The mean visits per episode for clinic states across both retail clinic and community episodes is 1.147. The P value compared to the non–clinic state mean of 1.430 is 0.657.

Relative episode costs

The allowed cost of treatment of AOM episodes was compared in the community and clinic settings. Table 4 shows an important result regarding the comparability of the retail clinic and non–retail clinic state claimants: the risk scores are not significantly different in any years. This is not true for the average risk scores of retail clinic patients, which are generally lower than those of community provider patients (in retail clinic states) as well as those of patients in non–clinic states. Thus, the retail clinics serve less risky patients within a claimant population that does not differ significantly from the claimant population in non–retail clinic states.

Table 4.

Average Risk Score for Claimants in Clinic and Non–Clinic States

		Retail Clinic States
Calendar Year		Retail Claimants	Community Claimants	Total Claimants^**	Non–Retail Clinic States	P
2009	No. Claimants	15	379	394	171
	Average Risk Score^*	0.504	1.196	1.169	1.276	0.713
2010	No. Claimants	112	883	981	481
	Average Risk Score^*	1.130	1.128	1.096	1.301	0.283
2011	No. Claimants	250	792	1025	459
	Average Risk Score^*	0.792	1.112	1.032	1.093	0.719
2012	No. Claimants	146	421	566	309
	Average Risk Score^*	0.607	1.107	0.978	1.186	0.400

Risk scores are member-month weighted averages for those eligible members with a recorded visit for acute otitis media in the specified plan year (claimants). Excludes claimants for retail clinic episodes in non–retail clinic states because of very low volume. The number of patients is lower than the totals in Table 3 because some claimants had more than 1 episode in given year.

Total claimants in retail clinic states do not always equal the sum of the retail clinic episodes claimants and community episode claimants because a claimant may have had more than 1 type of episode. When the episode type is shown separately, the claimant is included in each type. When the episode types are combined, the claimant is counted only once.

The study team calculated the cost per episode of AOM by year, both on an unadjusted and risk-adjusted basis (Figures 1 and 2). The bars represent the range of the 95^th confidence interval. In these figures, retail clinic states and episodes that are initiated in retail clinics are referred to as “RC.”

FIG. 1.

Cost per episode by year and place of service.

FIG. 2.

Cost per episode (risk adjusted).

On a risk-adjusted basis, P values for the difference between retail clinic and community episode costs in retail clinic states are significant (P < 0.05) in all cases except 2009. Note, however, that the number of retail clinic episodes in 2009 is small (n = 16).

Antibiotic prescriptions fills

The study team calculated the number of antibiotic prescriptions filled within 3 days of a visit for the clinic and non–clinic episodes (space does not permit these data to be shown; the data are available from the corresponding author on request). Data were not available to determine the frequency of antibiotic prescriptions written, but the team was able to assess the frequency of prescriptions filled within 3 days of a visit with an AOM diagnosis. A relatively small proportion of the population fails to fill a prescription; this percentage declines as the number of visits increases. Overall, 16.6% of patients fail to fill a prescription after a single visit, although this percentage is only 4.7% for retail clinic patients, possibly because of the convenient proximity of the pharmacist.

The percentage of fills in retail clinics is higher than that in the community, both in retail clinic and non–retail clinic states. It was found that, overall, 85.1% of patients filled 1 or more prescriptions within 3 days of a visit. This percentage varies between clinic patients (95.4%) and community patients (82.8% for community patients in states with retail clinics and 85.5% in non–retail clinic states). Patients with multiple visits were more likely to fill a prescription (among all patients, 83.4% with a single visit filled a prescription vs. 92.1% for patients with 3 or more visits).

Discussion

This study had 3 objectives:

1. Analyze the relative cost of AOM episodes in retail clinic and community settings. Figure 1 compares cost per episode for each setting for years 2009–2012. In unadjusted terms the retail clinic is the least costly, with little difference in cost between community costs in retail clinic and non–retail clinic states (except in 2012). Adjusting for the relative risk of the different populations, the difference between retail clinic and community episode costs in retail clinic states is broadly consistent with prior studies, and varies between $30 (2009) and $130 (2010), as shown in Figure 2.

2. The substitution vs. duplication hypothesis is a major concern of payers when considering the addition of retail clinics to a preferred network. The study team compared the frequency of episodes in states with retail clinic access to those in states without such access and found lower utilization/1000 in the states with retail clinic access. The team also examined follow-up community and ED services in episodes that begin in retail clinics. The proportion of episodes that began in the retail clinic but later migrated to the community was low in all years, with no evidence of an episode that began in the retail clinic later including an ED visit. On balance, the data do not appear to support the hypothesis that the retail clinic is generating overutilization.

3. Frequency of antibiotic prescriptions in retail clinic and community settings: data were not available to determine the frequency of antibiotic prescriptions written, but the study team was able to assess the frequency of prescriptions filled within 3 days of a visit with an AOM diagnosis. The percentage of fills in retail clinics is higher than that in the community, both in retail clinic states and in non–retail clinic states. The higher retail clinic fill percentage is consistent with percentages quoted in the literature: the most recent Cochrane review of antibiotics for AOM states that “Antibiotic use for AOM varies from 56% in the Netherlands to 95% in the USA, Canada and Australia.”²⁷ However, the Dutch study by Akkerman et al²⁸ from which this percentage is taken finds evidence of both underprescribing (11% of encounters) as well as overprescribing (18% of encounters) according to guidelines. Froom et al²⁹ compared prescribing behavior for physicians in the United States, United Kingdom, and the Netherlands; 97% of children in the United States were prescribed an antibiotic; 99% of children in the United Kingdom, and only 17% in the Netherlands were prescribed an antibiotic. The recent study by Mehrotra evaluates antibiotic prescribing for a range of acute respiratory infections (ARIs) and finds that retail clinics have higher adjusted antibiotic prescribing rates (95%) than primary care practices (85%) for antibiotics-may-be-appropriate ARI visits (including otitis media).³⁰ Lower community fill rates could be related to a number of factors: fewer prescriptions being written, failure to fill the prescription (the American College of Preventive Medicine reports that only between 50% and 70% of prescriptions are filled³¹), or spontaneous relief on the part of the patient.

Limitations

This study examines the behavior of a single employer population who had an incentive to use retail clinics, where these were available. The retail clinic state and non–retail clinic state populations are similar in terms of sex distribution but they differ slightly in terms of age band distribution. There also may be other differences that cannot be accounted for because of unobserved variables (for example, there were no detailed age/date of birth data, and, consistent with administrative claims data sets, no data on household income or race). No adjustments were made for unknown differences in contracted costs in different states. A lack of clinical information does not allow the study team to assess the severity of a specific AOM episode, although, like other authors, the team adjusts the episode costs for the relative risk profiles of the claimant populations. Finally, the difference in fill rates between patients of community providers and retail clinics is statistically significant and warrants further study to determine whether this is because of episode characteristics, such as underlying co-diagnoses as part of the AOM visit, prescribing patterns, or patients being more likely to fill at the convenient pharmacy.

Conclusion

With the increasing popularity of retail clinics, it is important for payers and population health professionals to understand the cost and quality outcomes of these providers.

This study examined the experience of patients with AOM treated in retail clinics and by community providers in states with retail clinics, and compared this experience to patients from the same employer in states without retail clinics. This study confirms results of earlier studies that retail clinics are a less-costly alternative to community settings for the treatment of episodes of AOM. There also is little evidence that retail clinics lead to duplication of services (patients receiving follow-up care in other settings), a frequent concern of critics of retail clinics. Although cost per treatment episode was significantly lower in the retail clinic, this was in part because of the lower risk profile of the retail clinic patients; adjusting for relative risk, the retail clinic is still lower cost. Prior studies have found differences in cost between $51–$55 (Thygeson et al) and $262 (Sussman et al, although this difference is an annual all-cost difference, not an episode difference). The present study difference on a risk-adjusted basis of between $30 and $130, depending on year, is consistent with these studies and confirms the fact that the retail clinic is a lower-cost provider of care, even adjusting for the lower risk of the patient population. The study team addressed the concern that retail clinic services duplicate those in the community rather than being substitutive by looking at the frequency with which episodes that begin in the retail clinic transfer to the community. This frequency is low in all years. The average number of visits per episode is similar in retail clinic and non–retail clinic states and the cost of the retail clinic episode remains lower than that of an episode treated in the community. Finally, although higher than those in the community, retail clinic antibiotic fill rates are consistent with other published studies of AOM patients.

Conflict of Interest

Mr. Duncan, Ms. Clark, and Ms. Wang declare the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: All authors are current or former employees of Walgreen Co.

Walgreen Co. funded the data acquisition, study development, analysis, and publication costs. Mr. Duncan was a paid consultant after his employment at Walgreen Co.

References

Mehrotra

, Liu

, Adams

, et al. The costs and quality of care for three common illnesses at retail clinics as compared to other medical settings. Ann Intern Med. 2009; 151:321–328.

Accenture.. Retail medical clinics: from foe to friend?. https://www.accenture.com/_acnmedia/Accenture/Conversion-Assets/DotCom/Documents/Global/PDF/Dualpub_21/Accenture-Retail-Medical-Clinics-From-Foe-to-Friend.pdf. Accessed November 9, 2015 .

Convenient Care Association.. Fact Sheet. Convenient Care Clinics: Increasing Access. http://www.ccaclinics.org/images/Documents/CCA_IncreasingAccess_2015.pdf. Accessed November 4, 2015 .

Sussman

, Dunham

, Snower

, et al. Retail clinic utilization associated with lower total cost of care. Am J Manag Care. 2013; 19:e148–e157.

Rohrer

, Angstman

, Bartel

. Impact of retail medicine on standard costs in primary care: a semiparametric analysis. Popul Health Manag. 2009; 12:333–335.

Kochar

. Comparing costs and quality of care at retail clinics with those at other medical settings. Ann Intern Med. 2009; 152:266–267.

Finkelstein

, Stille

, Nordin

, et al. Reduction in antibiotic use among US children, 1996–2000. Pediatrics. 2003; 112(3 pt 1):620–627.

US Interagency Task Force on Antimicrobial Resistance.. A public health action plan to combat antimicrobial resistance. 2001. http://www.cdc.gov/drugresistance/actionplan/aractionplan-archived.pdf. Accessed November 4, 2015 .

Schwartz

, Bell

, Hughes

. Preventing the emergence of antimicrobial resistance. A call for action by clinicians, public health officials, and patients. JAMA. 1997; 278:944–945.

10.

Klein

. The burden of otitis media. Vaccine. 2000; 19:S2–S8.

11.

Teele

, Klein

, Rosner

. Epidemiology of otitis media during the first seven years of life in children in greater Boston: a prospective, cohort study. J Infect Dis. 1989; 160:83–94.

12.

Vergison

, Dagan

, Arguedas

, et al. Otitis media and its consequences: beyond the earache. Lancet Infect Dis. 2010; 10:195–203.

13.

Bondy

, Berman

, Glazner

, Lezotte

. Direct expenditures related to otitis media diagnoses: extrapolations from a pediatric medicaid cohort. Pediatrics. 2000; 105(6):E72.

14.

Gates

. Cost-effectiveness considerations in otitis media treatment. Otolaryngol Head Neck Surg. 1996; 114:525–530.

15.

Ahmed

, Battacharyya

, Shapiro

. Incremental health care analysis of AOM. Laryngoscope. 2014; 124:301–305.

16.

Rosenfeld

, Kay

. Natural history of untreated otitis media. In: Rosenfeld

, Bluestone

, eds. Evidence-Based Otitis Media, Second Edition. Hamilton, Ontario: BC Decker, Inc.; 2003:180–198.

17.

Baxter

. Otitis media in Inuit children in the eastern Canadian arctic—an overview—1968 to date. Int J Pediatr Otorhinolaryngol. 1999; 49:S165–S168.

18.

O'Neill

, Roberts

, Bradley-Stevenson

. Otitis media in children (acute). BMJ Clin Evid. 2007; 2007: pii 0301.

19.

Finkelstein

, Metlay

, Davis

, Rifas-Shiman

, Dowell

, Platt

. Antimicrobial use in defined populations of infants and young children. Arch Pediatr Adolesc Med. 2000; 154:395–400.

20.

McCaig

, Besser

, Hughes

. Trends in antimicrobial prescribing rates for children and adolescents. JAMA. 2002; 287:3096–3102.

21.

McCaig

, Hughes

. Trends in antimicrobial drug prescribing among office-based physicians in the United States. JAMA. 1995; 273:214–219.

22.

Thygeson

, Van Vorst

, Maciosek

, Soberg

. Use and costs of care in retail clinics vs. tradional care sites. Health Aff (Millwood). 2008; 27:1283–1292.

23.

University of California San Diego.. Chronic Illness and Disability Payment System. http://cdps.ucsd.edu/. Accessed November 4, 2015 .

24.

Winkleman

, Mehmud

. A Comparative Analysis of Claims-based Tools for Health Risk Assessment. Schaumburg, IL: Society of Actuaries; 2008.

25.

Duncan

. Healthcare Risk Adjustment and Predictive Modeling. Winsted CT: Actex Publications; 2011.

26.

Angstman

, Bernard

, Rohrer

, Garrison

, MacLaughlin

. Repeat retail clinic visits: impact of insurance coverage and age of patient. Popul Health Manag. 2012; 15:358–361.

27.

Venekamp

, Sanders

, Glasziou

, Del Mar

, Rovers

. Antibiotics for acute otitis media in children. Cochrane Database Syst Rev. 2013; 1:CD000219.

28.

Akkerman

, Kuyvenhoven

, van der Wouden

, Verhij

TJM

. Analysis of under- and over-prescribing of antibiotics in acute otitis media in general practice. J Antimicrob Chemother. 2005; 59:569–574.

29.

Froom

, Culpeper

, Green

, de Melker

. A cross-national study of acute otitis media: risk factors, severity and treatment at initial visit. J Am Board Fam Pract. 2001; 14:406–417.

30.

Mehrotra

, Gidengil

, Setodji

, Burns

, Linder

. Antibiotic prescribing for respiratory infections at retail clinics, physician practices, and emergency departments. Am J Manag Care. 2015; 21:294–302.

31.

American College of Preventive Medicine.. Medication adherence—improving health outcomes. 2014. http://www.acpm.org/?MedAdhereTTProviders. Accessed December 31, 2014 .

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.04 MB

0.03 MB

0.00 MB