Massachusetts has insurance rates similar to those projected under the Affordable Care Act, but many of the state's patients are insured through private insurance plans with high out-of-pocket costs. We aimed to explore the relationship between insurance type (private vs public) and delays in care due to cost, stratified by income. 

We conducted a study of English-speaking adults recruited from the waiting rooms of the emergency department or outpatient clinics of a large healthcare system in western Massachusetts. Our primary outcome was the association between insurance type and cost-related delay in care, stratified by income. 

Of 800 individuals approached, 619 (77%) completed the survey. Participants were 60.6% male and 40.2% white, 37.2% Hispanic, and 12.6% black. The majority (61.4%) of those surveyed had public insurance, 34.1% had private insurance, and 4.5% were uninsured. Overall, 13.3% reported delays in seeking care that were related to cost. The impact of insurance on delay of care differed significantly by income tertile (P = .02): in the middle-income group ($12,500 to <$25,000 per person annually), privately insured respondents were more likely to delay care due to cost compared with publicly insured subjects (15.6% vs 8.1%; odds ratio [OR], 4.4; 95% confidence interval [CI], 1.9-10.2, unadjusted; OR, 2.2; 95% CI, 0.9-5.8, adjusted). 

Cost-related delays in care are prevalent despite the presence of an insurance mandate. Middle-income, privately insured patients report more cost-related delays in care compared with publicly insured patients with similar incomes. 

We surveyed patients and visitors in a large emergency department and outpatient clinic about type of insurance, out-of-pocket costs, and self-reported cost-related delays in receiving medical treatment. Middle-income respondents with private insurance reported more cost-related delays in care than middle-income respondents with public insurance, although adjustment attenuated these results. 

These findings suggest that national healthcare policies that have pushed middle-income patients toward high-deductible healthcare plans with high out-of-pocket costs may have the unintended consequence of increasing cost-related delays in care for this group.

Due to the worse health outcomes of the uninsured, the Affordable Care Act (ACA) aimed to expand health insurance coverage through multiple interventions and mandates: expanding Medicaid coverage; incentivizing employers to provide health insurance; and, for those who do not have employer- or government-provided coverage, implementing an individual mandate for the purchase of private health insurance on a regulated health insurance marketplace.

 Decreasing the number of uninsured aimed to improve population health by decreasing financial barriers to healthcare.

 Patients newly insured under the ACA are a heterogeneous group that includes those with employer-sponsored insurance, those with Medicaid, and those who purchased private individual plans from state and federal exchanges.

 Because of this heterogeneity, there is a risk that some types of health insurance will not allow patients to fully access needed care. This concern is all the more relevant however, given that the most rapidly expanding insurance types under the ACA are private plans with high levels of cost sharing.

 &shy;&shy;&shy;Although these plans offer more affordable monthly premiums,

 they are likely to have higher co-pays and deductibles than public insurance,

 which may put some patients at risk for reduced access to care due to cost.

Since 2006, Massachusetts has had a mandate requiring individuals to carry health insurance, resulting in less than 5% of the population being uninsured.

 This creates insurance coverage that is similar to the coverage patterns expected nationally after full implementation of the ACA.

 We hypothesized that individuals in Massachusetts with private insurance would be more likely to experience cost-related delays in care than participants with public insurance, and that this association would be related to household income and co-pays. To test this hypothesis, we surveyed patients and visitors in a large health system in Massachusetts to explore the relationship between insurance type and delays in healthcare due to cost.

We identified a convenience sample of adults recruited from the waiting rooms of an emergency department (ED) and 2 outpatient primary care clinics in western Massachusetts. The ED serves a diverse population, with 110,000 patient visits annually. The outpatient clinics in the study health system serve primarily low-income patients, with 49,000 visits annually. We approached adults who were not visibly in distress in the waiting rooms of the ED and the clinics. After confirming that they were at least 18 years old and English-speaking, we invited patients and those accompanying them to participate in this study. The Baystate Medical Center Institutional Review Board approved the study protocol; informed consent was waived given that no protected health information was collected other than zip code. Data were collected between February 2014 and May 2014.

Using previously validated instruments as a guide,

 we developed a questionnaire to measure participants’ healthcare utilization and their reported impact of cost on utilization of healthcare (eAppendix A [

]). Demographic data consisted of age, sex, race, education, employment status, annual pre-tax household income, zip code, and household size.

Our exposure variable, insurance type, was categorized as uninsured, private insurance, or public insurance. Participants were also asked to specify whether they had an individual or a family health insurance plan. We further asked whether private insurance was self-purchased or employer-sponsored, as self-purchased insurance is more likely to have higher out-of-pocket (OOP) costs.

 Public insurance included Medicare and Medicaid (referred to as “MassHealth” in Massachusetts), as well as a subsidized public—private insurance partnership called “Commonwealth Care.” We categorized both Medicare and Medicaid as public insurance because both have much lower premiums and less cost sharing than private insurance (although they have different patient populations and payment methods).

Until January 30, 2015, Commonwealth Care was a state-subsidized plan available to individuals whose income was not low enough to qualify for Medicaid, but for whom private insurance at market rates would be a significant financial burden. Under the ACA, there is no clear comparator with Commonwealth Care, but as Commonwealth Care was highly subsidized by the state, with co-pays comparable with Medicaid’s in the lowest-income population, we opted to categorize it under public insurance.

 This decision was supported by the fact that Commonwealth Care respondents’ demographics were more similar to Medicaid/Medicare patient demographics than to the privately insured group.

The questionnaire also included questions about health and access to healthcare. We collected participants’ self-reported number of chronic medical conditions and state of health,

 as well as a variety of previously validated measures of access to care, including number of outpatient visits in the past year,

 For those who responded “yes” to delays of medical care, we asked additional questions about type of medical care that was delayed and the reasons for delaying care.

The primary outcome was delay in care due to cost, defined as voluntary delay or refusal of care by the respondent due to the cost of care involved. That is, we defined respondents as such if they affirmatively answered the question, “In the past year, have you ever delayed or avoided getting any kind of medical care?” and then selected the reason, “It cost too much money” when asked, “Why did you delay or avoid medical care?”

We first pilot-tested the survey with a sample similar to the target audience using cognitive interviewing techniques.

 The goals of pilot testing were to ensure clarity and completeness of the questions and to assess administration and completion time. We then revised the questionnaire based on the feedback we received. No data obtained from the pilot testing were included in the analysis. The survey administration details can be reviewed in eAppendix B.

We generated descriptive statistics as proportions for categorical factors and medians, with interquartile range (IQR) for ordinal variables. We next calculated per-person household income (PPHI) by dividing the median value of self-reported pre-tax household income category by the number of household residents. For respondents with missing PPHI, multiple imputation was used to estimate the value using available education, interview site (ED vs clinic), self-rated health, and self-rated OOP costs as predictors, chosen for their relatively high correlation with income in the complete sample.

 We then divided PPHI into tertiles. We included uninsured respondents in descriptive analyses and in calculations of income tertiles to ensure a representative sample. However, we excluded the uninsured from multivariable models as this group was too small to draw meaningful conclusions.

Because we hypothesized that those with lower incomes would be more sensitive to high co-pays and deductibles than those with higher income, we tested an “income-level by insurance category interaction term” using the Likelihood Ratio (LR) test. If the 

 value for the LR test of the interaction term was ≤0.2, a stratified approach would be used and separate estimates for the association between delay of care and insurance type would be presented for each income level. Final model fit was examined using the Hosmer-Lemeshow Goodness-of-Fit test. After multivariable adjustment, we wished to quantify the extent to which OOP costs and outstanding bills “explained” differences in delay of care that remained after multivariable adjustment. To do so, we added each variable to the fully adjusted model and noted the change in the insurance variable’s beta coefficient.

 We used Stata version 13.1 (StataCorp; College Station, Texas) for all analyses.

Of 800 individuals approached, 704 responded to the survey (88%). Of these, we excluded 85 (12%) questionnaires that were missing insurance category information or for which there was no response to the question regarding delayed care. Thus, we had 619 surveys completed of the 800 subjects originally approached, giving us a final response rate of 77%. Compared with those included in the study, the 85 excluded participants (who failed to complete the survey) were generally older (median age = 49.5 vs 39.5 years; 

= .02), less educated (median years of education 12 vs 14; 

= .001), had higher median comorbidity counts (7 vs 2; 

<.001), and were more likely to be surveyed in a clinic versus the ED (68.2% vs 23.4%; 

Of the 619 participant responses analyzed, 474 participants (77%) were recruited from the ED and 145 (23%) from 2 primary care clinics. Participants’ characteristics appear in Table 1.

Tertiles of the pre-tax PPHI distribution corresponded to cut points from the 2014 MassHealth Income Standards and Federal Poverty Guidelines

: (tertile 1: <$12,500; tertile 2: $12,500 to <$25,000; tertile 3: ≥$25,000). Household income (HHI) was missing for 111 respondents (17.9%), and we were able to impute HHI for all but 16, who were excluded from the multivariable analysis.

The median number of visits to a provider within the past 12 months was 3 (IQR = 2-8). Eight percent reported delays in filling prescriptions, and 32% reported outstanding medical bills. One-third of those analyzed reported OOP costs upwards of $500 annually (32.8%) (Table 1). Overall, 189 of the 619 subjects (30.5%; 95% confidence interval [CI], 27.0%-34.3%) reported any delay in care in the 12 months prior to completing the survey (data not shown). Of the 189 who delayed care for any reason, 82 (43%) did so due to cost. Other nonmutually exclusive reasons for delay are shown in Figure 1.

Forty-three percent had more than 1 chronic medical condition, with 9% reporting upwards of 5 chronic medical conditions for which they receive prescriptions. In terms of self-reported health, 35.4% reported their health to be very good or excellent and 25.7% reported their health to be fair or poor (Table 1).

The uninsured were similar to the publicly insured in terms of race and income level, but were younger overall (Table 1). The uninsured had higher OOP costs than those with public insurance but less than those with private insurance (Figure 2).

As shown in Table 1, demographic characteristics and annual OOP spending varied by insurance type. For example, compared with publicly insured participants, those with private insurance were more likely to be employed (76.8% vs 30.8%; 

 <.001) and to have family insurance plans (64.5% vs 26.6%; 

 <.001). They were also more likely to identify as white (62.6% vs 28.7%), have income greater than $100,000 per year (11.9% vs 1.6%), have a graduate degree (14.2% vs 4.7%), and to have been surveyed in the ED as opposed to the clinic (89.1% vs 68.9%). Compared with publicly insured participants, those with private insurance were less likely to report 5 or more chronic conditions requiring medication (2.8% vs 12.9%). Respondents with public insurance were significantly less likely than those with private insurance to have OOP expenditures greater than $500 per year (Figure 2).

Characteristics of respondents who delayed care due to cost are shown in the eAppendix Table. Delay in care due to cost was highly correlated with a delay in filling prescriptions (41.5% of those who delayed care avoided filling a prescription vs 3.4% who did not delay care; 

<.001), but was not significantly associated with frequent ED use for healthcare (17.1% vs 11.7%; 

 = .21). Delay of care due to cost was significantly associated with white race/ethnicity (whites composed 54.9% of those who delayed care vs 38.0% of those who did not), being employed (64.6% in the delay group vs 44.7% in the nondelay group), and having outstanding medical bills (72.0% vs 26.6%). Annual OOP costs less than $500 were associated with fewer reported delays in care (42.7% vs 67.8%).

In the multivariable model, examining delay of care by income level and insurance type, we found that the income tertile by insurance type interaction term was statistically significant (

= .02), indicating the need for a stratified approach. The association between delay of care and insurance type, by income stratum, is shown in Table 2. Table 2 also includes standardized effect sizes so that the magnitude of the difference between groups can be interpreted without respect to sample size,

 because the stratified approach resulted in small subgroups and multivariable adjustment necessitated exclusion of respondents with missing covariates.

In unadjusted models, privately insured subjects in the middle-income group were 4.4 times (95% CI, 1.9-10.2; 

= .001, unadjusted) more likely to delay care compared with publicly insured subjects. After adjustment, privately insured middle income patients were 2.2 times (95% CI, 0.9-5.8; 

= .10, adjusted) more likely to delay care due to cost compared with publicly insured subjects (adjusted proportions = 15.6% vs 8.1%; Cohen’s d, +0.4, moderate).

Although the association was no longer statistically significant after adjustment (95% CI, 0.9-5.8), we refer to the effect size rather than the statistical significance because multivariable adjustment (and exclusion of respondents with missing covariates) may have affected statistical power. Of note, lower-income, privately insured respondents were less likely than publicly insured respondents to delay care due to cost (adjusted proportions 3.1% vs 14.1%; odds ratio 0.2; 

= .21; Cohen’s d, —0.9, large), although this finding did not achieve statistical significance in adjusted or unadjusted models. We noted only very small differences in delay of care between private and public insurance in the highest income tertile.

After multivariable adjustment, we wished to quantify the extent to which OOP costs and outstanding bills “explained” differences in delay of care. In the lowest-income tertile, adjustment for OOP costs increased the protective effect of private insurance by 16%, whereas outstanding bills did so by 1%. In the middle-income group, adjustment for OOP costs removed 27% of the risk for delay of care for privately insured subjects, whereas outstanding bills removed 11%. This suggests that some of the observed association between delay of care in privately insured, middle-income respondents is due to higher OOP costs and more outstanding bills compared with their publicly insured counterparts.

In a large convenience sample in Massachusetts, we found that one-third of patients reported a delay in care in the year prior to the survey. About half of the time, this delay was due to cost. We also found that, after stratifying by income, middle-income tertile ($12,500 to <$25,000 PPHI) participants with private insurance were more likely to report cost-related delays in care compared with participants in the same income group with public insurance. Adjustment for demographic confounders—race, employment status, and self-rated health&mdash;attenuated but did not eliminate the association, as evidenced by a medium effect size even after multivariable adjustment. Annual OOP costs (eg, co-pays, deductibles) partially mediated the association between insurance type and delay due to cost.

Our findings may, at first, appear to contradict findings from a recent study conducted by McCormick et al that found Medicaid was associated with more delays in care compared with private insurance or Commonwealth Care.

 However, there are several key differences between their study and ours. First, they collected data only from patients in the ED, which might indicate that they were more likely to sample patients without a regular source of care. In contrast, we also sampled visitors in outpatient clinics, thus including those who were not directly using the healthcare system at the time the survey was completed. Second, they did not stratify by income, and, as we have noted, income appears to be an effect modifier in the relationship between insurance type and delays in care. Third, they did not include Medicare patients, meaning that their “Medicaid” group was different from our “public” insurance group. To assess the extent to which this contributed to the difference in our results, we conducted a sensitivity analysis by excluding patients 65 or older (eAppendix C). We found that results did not change appreciably, supporting the original model. Finally, we included Commonwealth Care under the publicly insured group, which makes the studies less comparable. It is notable that both studies show similar overall proportions of participants who delayed care.

 Another study, by Bernard et al, surveyed workers who declined employer-sponsored health insurance and found that those with public insurance had similar access to care compared with those with private insurance.

 Again, the authors did not stratify by income, which could be important since employed workers who declined insurance are likely to be a self-selected group. Notably, a study that did collect income data, by Wisk et al, reported that the healthcare-related financial burden in families with children was dependent on income.

A 2015 Gallup survey reported that, in the prior year, 31% of Americans delayed seeking healthcare because of cost.

 In contrast, we found that 13% of our sample delayed care because of cost in the year prior to the survey. We suspect that there are 2 reasons why we reported fewer cost-related delays in care than were identified in the Gallup survey. First, Massachusetts—with insurance rates above 90%&mdash;is more reflective of the post-ACA rather than pre-ACA world. Thus, our study may be an indicator of upcoming trends: if the ACA remains law, it is possible that, in the near future, we will see fewer overall delays in care, but an increasing number of middle-income patients with private insurance who are experiencing delays in care.

 Second, we surveyed subjects who were in a healthcare setting, either in an ED or an oupatient clinic waiting room. This population may be different than the randomly dialed national sample obtained by the Gallup poll.

There are several limitations to our study. First, we relied on an English-speaking convenience sample in a single geographical area—western Massachusetts&mdash;which may not be representative of the rest of the United States. We attempted to address this limitation by conducting surveys in an ED (which serves a large regional catchment area) and clinics that serve a lower-income clientele (who are primarily publicly insured). We included only individuals who were in the process of accessing medical care directly or indirectly (as visitors or companions), which is a population that may differ from the overall population. However, the demographic characteristics of our sample population were comparable with those of the state overall.

 Because the ACA was based on Massachusetts’ 2006 healthcare reforms, local studies, such as ours, can inform projections about the insurance landscape under the ACA in the rest of the country, assuming that the ACA remains intact in the future.

A second limitation of our study is the fact that insurance information was self-reported, preventing us from collecting details regarding insurance plans and coverage. For example, there were 3 types of Commonwealth Care that had different levels of cost sharing, but most patients were not aware of these differences. We attempted to address this limitation by asking separate questions about OOP costs. We also included Commonwealth Care, Medicare, and Medicaid in a single public insurance category. Another limitation is the fact that we collected income information in a categorical fashion and imputed some income data. Although categories were designed to improve response rates and increase respondent comfort with providing delicate information, their categorical nature limited our ability to assign PPHI that corresponded with federal poverty guidelines. Nevertheless, this method of reporting income has been reported to have acceptable accuracy.

Finally, given that the majority of patients sampled were in the ED, this can bias the overall population to those with less access to care (those with a usual provider would be less inclined to go to the ED for low-acuity issues). However, given that we not only sampled patients, but also family and friends, we hoped to include a population that was not directly accessing the healthcare system. Future studies should target outpatient clinics in areas with varied demographics to better delineate the effect of insurance type on access to care, with income as an effect modifier.

In spite of these limitations, our study also has important implications. Our study results confirm that patients with health insurance have more access to care than those without insurance. Thus, higher insurance rates achieved under the ACA will likely improve access for patients who were previously uninsured. However, our study also highlights that one-third of patients with health insurance still experience delays in care. Privately insured patients in our study had higher OOP costs than those without insurance, and we found that many people who were privately insured and living above the poverty line (the middle-income group) frequently experienced recent delays in care due to cost. This may be because this group has the means to purchase private insurance, but may not have sufficient disposable income to cover large unexpected medical bills.

A potential unintended consequence of an insurance mandate is that employers and patients will increasingly embrace plans with high rates of cost sharing, resulting in large numbers of privately insured patients who are unable to afford recommended care. To reduce the chance that this could happen, our study results indicate that policy makers should pay close attention to the regulation of insurance plans. In particular, efforts should be made to ensure robust minimum coverage and to maintain or expand current limitations on OOP expenses. Most importantly, our findings suggest that there are great risks to current political efforts aimed at loosening minimum healthcare coverage and reducing limitations on OOP medical expenses. If these efforts succeed and plans with even greater cost sharing are allowed to flood the market, the result could be widespread increases in cost-related delays in care in susceptible sociodemographic groups, specifically middle-income beneficiaries with private insurance.

Proposed changes to healthcare law (eg, the repeal of the ACA and replacement with an alternative plan) would likely result in a smaller proportion of the population who can obtain public insurance (because of cuts to Medicaid), and, for those with private insurance, a loss of guaranteed minimum healthcare coverage and fewer limitations on OOP spending. This study is an important addition to the ongoing policy debate, as the results suggest that these changes would result in more cost-related delays in access to care for middle-income populations with private insurance (compared with similar populations with public insurance). Policy makers considering such changes to current law should be aware that higher OOP costs are likely to increase delays in access to healthcare, especially for populations with private insurance.&ensp;

 Department of Internal Medicine, Divison of Primary Care, Tufts Medical Center (SAR), Boston, MA; Center for Value-Based Care Research, Medicine Institute, Cleveland Clinic (MBR), Cleveland, OH; Department of Internal Medicine, Gallup Indian Medical Center (BJ), Springfield, Gallup, NM; Department of Internal Medicine, Berkshire Medical Center (MAM), Pittsfield, MA; Cigna HealthCare (JF), Bloomfield, CT; Department of Emergency Medicine, San Jose Regional Medical Center (JM), San Jose, CA; Center for Quality of Care Research, Department of Medicine, Tufts University School of Medicine (SLG, TL), Boston, MA; Baystate-University of Massachusetts Medical School (TL), Springfield, MA.

 Dr Lagu is supported by the National Heart, Lung and Blood Institute of the National Institutes of Health under Award Number K01HL114745.

 The authors report no relationship or financial interest with any entity that would pose a conflict of interest with the subject matter of this article.

Articles

POLICY

In a survey of patients and visitors to a large academic medical center, middle-income respondents with private insurance reported more cost-related delays in care than those with public insurance.

The Association Between Insurance Type and Cost-Related Delay in Care: A Survey

 To measure variations in drug prices across and within zip codes that may reveal simple strategies to improve patients’ access to prescribed medications.

We compared drug prices at different types of pharmacies across and within local markets. In-store prices were compared with a Web-based service providing discount coupons for prescription medications. Prices were collected for 2 generic antibiotics because most patients have limited experience with them and are less likely to know the price ranges for them.

Drug prices were obtained via telephone from 528 pharmacies in Los Angeles (LA) County, California, from July to August 2014. Online prices were collected from GoodRx, a popular Web-based service that aggregates available discounts and directly negotiates with retail outlets.

Drug prices found at independent pharmacies and by using discount coupons available online were lower on average than at grocery, big-box, or chain drug stores for 2 widely prescribed antibiotics. The lowest-price prescription was offered at a grocery, big-box, or chain drug store in 6% of zip codes within the LA County area. Drug prices varied dramatically within a zip code, however, and were less expensive in lower-income areas. The average price difference within a zip code was $52 for levofloxacin and $17 for azithromycin.

Price shopping for medications within a small geographic area can yield considerable cost savings for the uninsured and consumers in high-deductible health plans with high negotiated prices. Clinicians and patient advocates have an incentive to convey this information to patients to improve adherence to prescribed medicines and lower the financial burden of purchasing prescription drugs.

The high costs of prescription drugs and lack of price transparency suggest that many patients might be overpaying for medications. We found that: 

Drug prices were lowest at independent pharmacies and when using online coupons compared with grocery, big-box, or chain drug stores. 

The average price difference within a zip code was $52 for levofloxacin and $17 for azithromycin. 

Medications were less expensive when purchased in lower-income areas. 

Clinicians and patient advocates have an incentive to convey this information to their patients or members to improve adherence to prescribed medicines and lower the financial burden of purchasing prescription drugs. 

Consumers typically know the price of a product and have some information about its quality before purchasing. This information makes comparison shopping possible and is a key tenet of well-functioning markets. However, healthcare markets are different. Patients rarely know the price of a medical product or service before using it and sometimes even after the service is provided.

A lack of price transparency and difficulty in assessing its quality are the likely key reasons the price of medical care varies so widely. A primary care doctor visit in San Francisco ($251) is twice as expensive as in Miami ($95), and a lipid panel in Pittsburgh ($19) is one-fourth of the price in Indianapolis ($89), which is just a fraction of the cost in Dallas ($343). Even within the same market, New Yorkers seeking magnetic resonance imaging (MRI) can pay anywhere between $416 and $4527 for the same service.

Although millions of Americans gained coverage through the Affordable Care Act (ACA), more than 32 million consumers remain uninsured and exposed to the full cost of services.

 Many of the uninsured are low-income consumers, who lack coverage as a result of states’ decisions not to expand Medicaid under the ACA. In addition, an increasing number of Americans are enrolled in health plans with high annual deductibles and face potentially high out-of-pocket costs. Four of 5 workers who now receive insurance through an employer pay a deductible, and 1 in 5 faces a deductible of $2000 or more.

 Plan members can purchase covered medications at a negotiated rate that might not vary much across in-network pharmacies; however, the plan’s negotiated price can be significantly greater than what consumers would pay using a store’s proprietary discount card or with online coupons. As such, both the uninsured and an increasing number of insured consumers have incentives to use price information and discount programs to comparison shop for their prescription medications.

Several states have passed laws and private-sector initiatives are underway to encourage or require greater price transparency for medical services.

 Recent work suggests that providing employees with price information is associated with lower total claims payments for common medical services.

 In this paper, we examined the potential benefit of comparison shopping for prescription drug prices and pharmacy characteristics that patients should consider when shopping for medications. Prescription drugs provide a good test case because, unlike most medical services, the quality of the product is constant across providers, making it easy for patients to comparison shop. Further, the market for prescription drugs is generally local: the majority of Americans live within 5 miles of a pharmacy and about 80% of the population uses only 1 pharmacy for outpatient prescriptions.

 We compared drug prices from different types of pharmacies (ie, chain drug stores, independent pharmacies, grocery stores, big-box stores) within local markets and described how they compare with the prices of Web-based services offering discount coupons for prescription medications. We measured the extent to which prices vary within a zip code and whether drug prices vary in high- versus low-income areas. The extent of price variation within a market is an implicit measure of the benefit of price shopping and may reveal simple strategies that health plans and providers can use to help patients access prescribed medications. 

We conducted a cross-sectional analysis of drug prices from July to August 2014 from all outpatient prescription drug outlets in a selected set of zip codes in Los Angeles (LA) County. We focused on the prices of 2 commonly prescribed generic antibiotics for community-acquired pneumonia: levofloxacin and azithromycin. They are relatively equally prescribed and can be used as substitutes for each other in most areas. We purposely selected medications used for an acute medical condition because patients typically have limited experience with them and are less likely to know the range of prices for these products. Further, the consequences of not filling medications for acute infections, such as pneumonia, may lead to sepsis, resulting in otherwise preventable hospitalizations, increased healthcare costs, and greater risk of mortality. The acute nature of their use and limited time to price shop after diagnosis may result in greater price variation than for chronic medications, although there is limited evidence on this point.

We used the American Community Survey to obtain zip code—level measures of population and median household income for LA County. As low-density areas have few pharmacies, we restricted the study sample to zip codes with at least 10,000 residents. We ranked the 164 remaining zip codes by median household income and selected the top (high-income) and bottom (low-income) quartiles for inclusion.

We obtained a complete list of pharmacies operating in these 82 zip codes from the California Board of Pharmacy. We excluded hospital-based pharmacies and membership clubs (eg, Sam’s Club, Costco) and categorized outpatient pharmacies in the following way: chain drug store (eg, CVS), independent pharmacy, grocery store (eg, Safeway), or big-box store (eg, Target). Prices at these pharmacies were obtained via telephone. Three trained research assistants (RAs) called all pharmacies on the list over a 1-month period from July to August 2014. They followed a standardized script informing pharmacy staff that they were calling from LA County Hospital on behalf of a hypothetical uninsured patient with pneumonia.

They then requested quotes for the cash prices for the generic forms of 7 tablets of 500-mg levofloxacin and 6 tablets of 250-mg azithromycin. Once a price was given, they asked about any available discounts that could lower the price of the drug (ie, “discounted price.”) If there was no answer, the pharmacy staff was too busy to run a price check, or the RA was put on hold for more than 10 minutes, the RA would call back every hour with the same request until 5 pm that day and then resume hourly calls the next day at 8 am. Phone prices for both medications were recorded on a standardized data collection sheet.

Given the growth in internet use and online purchases of prescription drugs, we simultaneously collected prices for the 2 medications obtained at the same set of pharmacies from GoodRx, a popular Web-based service that aggregates available discounts and directly negotiates with retail outlets to provide consumers with coupons for discounted drug prices. Patients can enter a medication name and zip code and the website will list prices at most pharmacies operating in or near that zip code. To test the acceptance of the GoodRx coupons, we physically presented them at 5% of the pharmacies to ensure their prices would be honored. The study protocol was reviewed by the University of Southern California’s institutional review board prior to initiation. We examined the distribution of discounted drug prices by pharmacy type, as well as the extent of price variation in high- versus low-income areas, and then explored how prices varied across pharmacies in the same zip code. Given that most individuals purchase medications near home, examining prices within a zip code is an implicit measure of the potential cost savings from price shopping.

We obtained drug prices from 528 of 535 eligible pharmacies (98.7% response rate). The sample consisted of 170 chain drug stores, 49 grocery stores, 39 big-box stores, and 270 independent pharmacies. Independent pharmacies accounted for a larger proportion of the total number of pharmacies in low-income zip codes (65%) than in high-income areas (37%).

 shows the variation in drug prices by pharmacy type, categorized as chain, independent, grocery store, big-box, and online (GoodRx). This captures variation in prices both across and within zip codes and reflects the discounted price. The average price of generic levofloxacin purchased with a GoodRx coupon or at an independent pharmacy was less than half the price versus a grocery or big-box store and less than one-fourth of the discounted price obtained over the phone at chain drug stores. Although prices were highest at chain drug stores, there was far less variation in price at this type of location (interquartile ratio [IQR], 1.05) compared with the other retail outlets (IQRs, ~2.0).

Prices varied less for azithromycin than levofloxacin, but relative prices followed a similar pattern. The lowest average prices were found via GoodRx ($20) and at independent pharmacies ($23); chain drug stores charged the most ($37). There was little variation in price for chain drug stores (IQR, 1.05), particularly in comparison with independent pharmacies (IQR, 2.47) and grocery stores (IQR, 4.30). Asking for a discount had the largest effect at chain drug stores in the case of levofloxacin (lowering the average price by $11, or 10%) and at grocery stores in the case of azithromycin (by $8, or about 25%).

 shows the distribution of drug prices in high- and low-socioeconomic status (SES) areas, as defined by the median household income in the zip code. Levofloxacin and azithromycin were less expensive, on average, when purchased in pharmacies located in lower-income zip codes compared with higher-income zip codes. Because high- and low-income areas are likely to have a different mix of pharmacies (eg, big-box stores may be less likely to locate in low-income areas), we also estimated the average prices controlling for pharmacy type and the results were substantively unchanged. Although the median price of levofloxacin was $26 in low-income areas, prices ranged from a low of $4 to a high of $149, depending on the pharmacy. In high-income areas, the median price of levofloxacin was $80 and ranged from $5 to $229.

 shows the extent of price variation within the same zip code, which is the implicit value of price shopping within a localized area. The average price difference between the highest and lowest-cost pharmacies in a zip code was greater than $100 for levofloxacin and $30 for azithromycin. Perhaps a more salient comparison is the average price difference between a randomly selected pharmacy and the lowest-priced pharmacy in the same zip code. In this case, consumers would save an average $52 per prescription for levofloxacin and $17 per prescription for azithromycin simply by comparison shopping within the same zip code (results not shown). We observed modestly greater price variation in high-income areas despite there being more pharmacies per zip code in low-income areas (7.7 pharmacies per zip code vs 6.5 in high-income zips).

 highlights general approaches for obtaining the lowest priced medication in an area. In more than half of the 71 unique zip codes in the study sample, independent pharmacies had the lowest price for levofloxacin (53%), followed by GoodRx (44%). In only 2 of 71 zip codes did a chain or big-box store have the lowest price for levofloxacin. We observed the same pattern when we restricted the analysis to the 39 zip codes with a big-box store. For both levofloxacin and azithromycin, the lowest price prescription was offered at a grocery, big-box, or chain drug store in just 6% of zip codes.

In our study sample of 528 pharmacies, prices found at independent pharmacies and by using online discount coupons were markedly lower, on average, than at grocery, big-box, or chain drug stores for 2 widely prescribed antibiotics. Drug prices varied dramatically within a zip code and typically were less expensive when purchased in lower-income areas. The average price difference within a zip code was $52 for levofloxacin and $17 for azithromycin, which suggests that price shopping within a small geographic area can yield considerable cost savings, particularly for uninsured and insured consumers in high-deductible health plans with high negotiated prices. A possible explanation for the greater price variation with levofloxacin is that it recently became available as a generic, so there has been less time to establish a fair market value.

There is a common perception that chain drug stores have lower prices than independent pharmacies due to economies of scale and the fact that the chains derive a smaller fraction of their revenue from the sale of prescription drugs. However, chain drug stores typically compete less on price and more on convenience, brand name, and nondrug items.

 By contrast, independent pharmacies compete largely on price and service to induce consumers to bypass chain drug stores.

 Our results suggest that cash-paying consumers often face a premium for going to chain drug stores and could save substantially by using online coupons or purchasing their medications at independent pharmacies in the same or neighboring zip codes.

Although poor adherence is endemic, it is particularly problematic for individuals of lower SES.

 An estimated 20% to 35% of patients are primary nonadherent by failing to fill an initial prescription, and an additional 20% discontinue therapy after filling the first prescription.

 In 2012, 22% of uninsured adults aged 18 to 64 years reported not getting needed prescription drugs due to cost compared with 5% of adults who were insured for the whole year.

 Noncompliance with antibiotics for respiratory infections can result in treatment failure, worsening severity of disease, sepsis, antibiotic resistance, and increased risk of hospitalization.

A critical question is the extent to which consumers would use price information in purchasing medical services. The highly touted movement toward “consumer-directed healthcare” relies on patients having easy access to information concerning drug prices and quality. A recent survey indicates that a majority of Americans have tried to find out how much they would have to pay out of pocket—not including a co-pay&mdash;before getting care. However, the survey also reveals that most Americans are not aware that prices can vary across healthcare providers.

Our results differ from those of a Florida study by Gellad et al that obtained drug prices for 3 chronic medications (esomeprazole, fluticasone, and clopidogrel) and a generic antibiotic (azithromycin).

 They found that mean drug prices were 9% higher in the poorest  zip codes and that independent pharmacies in the poorest areas charged the highest mean prices. We, however, found the opposite: lower prices at independent and online pharmacies and pharmacies located in low-income areas. A possible explanation for the differences across studies is that the Florida study obtained drug prices from a website whereas we collected prices by calling individual pharmacies. We also asked for any available discounts and verified concordance with in-store prices in a pilot study. The Florida study also restricted the sample to pharmacies that filled 1 of the 4 drugs to a Medicaid beneficiary in a single month (November 2006). This may have resulted in a nonrepresentative sample of pharmacies across income areas. By contrast, we collected price data over the phone from all available pharmacies. We focused on price variation for antibiotics under the assumption that consumers would have limited experience purchasing them (and thus would be less aware of price) and the consequences of not filling the prescription due to cost would have a more immediate impact on health.

There are several limitations to our study, the most prominent of which is that we only measured prices of 2 medications for an acute condition in a single county in California (LA County). We do not know if the findings will hold in other regions or states. However, LA is an economically and culturally diverse county with a broad array of income levels and population densities. The magnitude of price variation across outlets and the savings associated with online coupons at nationwide chains suggest that we could expect similar results in other areas of the country. The extent of price variation may be lower for chronic medications, but these by their nature (length of time taken) may impose a larger financial burden on patients. Future research should focus on comparing prices across a broader spectrum of pharmaceuticals, including medications for chronic diseases.

Another limitation of our study is that we obtained drug prices via telephone rather than in person, and pharmacies may offer discounts in the store that they are unable or unwilling to provide over the phone. Also, the calls to the pharmacy were made from a doctor’s office on behalf of a hypothetical uninsured patient, and the callers asked each pharmacy for any potential discounts after an original price was provided. Over 98% of pharmacies in our sample provided prices over the phone. Patients calling on their own behalf may not receive the same discounts we received. Additionally, we only called pharmacies in the highest and lowest quartiles of median income. It is possible that we might have a better understanding of price variation if we had contacted all pharmacies regardless of income level.

Finally, we used a single website to represent discounts available online. Nonetheless, GoodRx is the largest price aggregator and coupon tool used by thousands of doctors and millions of patients every month. Further, 100% of GoodRx coupons were honored when physically presented at the pharmacy during this study.

Slowing the growth of healthcare costs underscores nearly every health policy initiative in the United States and is the motivation for public and private efforts to increase price transparency in healthcare markets. Price transparency initiatives face considerable obstacles, however; most prominently, how to reliably measure and convey information about quality and price for thousands of complex medical services produced by a wide array of providers and organizations. The task is less daunting for prescription drugs because quality is fixed.

The extent of price variation found in this study suggests that consumers could readily benefit from greater price transparency. If this information were widely available to consumers, large variations in drug prices across pharmacies would likely be reduced.&ensp;

The authors would like to thank Kaleigh Barnes, Brian Raffetto, Melissa Luttio, Janice Rivelle, and Erin Higginbotham for helping to collect and analyze the data.

 Keck School of Medicine, (SA, ST); Sol Price School of Public Policy (NS); and School of Pharmacy (GJ), Leonard D. Schaeffer Center for Health Policy and Economics, University of Southern California, Los Angeles, CA.

 Concept and design (SA, GFJ, NS, ST); acquisition of data (NS, ST); analysis and interpretation of data (SA, NS, ST); drafting of the manuscript (SA, GFJ, SN, ST); critical revision of the manuscript for important intellectual content (SA, GFJ, NS, ST); statistical analysis (SN); administrative, technical, or logistic support (ST); and supervision (SA, ST).

 Geoffrey F. Joyce, PhD, Leonard D. Schaeffer Center for Health Policy and Economics and School of Pharmacy, University of Southern California, 635 Downey Way, Los Angeles, CA 90089-3333. E-mail: gjoyce@usc.edu. 

1. Costliest cities 2015: pricing for the same medical services is all over themap (literally). Castlight health website. www.castlighthealth.com/costliest-cities/. Accessed September 27, 2015.

2. Garfield R, Majerol M, Damico A, Foutz J. The uninsured: a primer—key facts about the health insurance and the uninsured in the era of health reform. Kaiser Family Foundation website. http://files.kff.org/attachment/primer-the-uninsured-a-primer-key-facts-about-health-insurance-and-the-uninsured-in-the-era-of-health-reform. Published November 2015. Accessed January 7, 2016.

3. 2015 Employer Health Benefits Survey. Kaiser Family Foundation website. www.kff.org/report-section/ehbs-2015-section-one-cost-of-health-insurance/. Published September 22, 2015. Accessed September 27, 2015.

4. Kullgren, JT, Duey, KA, Werner, RM. A census of state health care price transparency websites. 

. 2013;309(23):2437-2438. doi: 10.1001/jama.2013.6557.

5. Whaley C, Schneider Chafen J, Pinkard S, et al. Availability of health service prices and payments for these services. 

. 2014;312(16):1670-1676. doi: 10.1001/jama.2014.13373.

6. Look KA. Patient characteristics associated with multiple pharmacy use in the U.S. population: findings from the Medical Expenditure Panel Survey. 

. 2015;11(4):507-516. doi: 10.1016/j.sapharm.2014.10.004.

7. Gellad WF, Choudhry NK, Friedberg MW, Brookhart MA, Haas JS, Shrank WH

 Variation in drug prices at pharmacies: are prices higher in poorer areas?

. 2009;44(2, pt 1):606-617. doi: 10.1111/j.1475-6773.2008.00917.x.

8. Four big trends shifting the business of retail pharmacy. McKesson website. http://www.mckesson.com/blog/retail-pharmacy-trends-to-watch/. Published March 14, 2016. Accessed June 2017.

9. Leger M. Three ways small drug stores can beat big-box competitors. The Globe and Mail website. https://www.theglobeandmail.com/report-on-business/small-business/sb-marketing/three-ways-boutique-drug-stores-stay-competitive/article12468461/. Published June 14, 2013. Accessed June 2017.

10. Goldman DP, Smith JP. Can patient self-management help explain the SES health gradient? 

CLINICAL

Price shopping for medications within a small geographic area can yield considerable cost savings for uninsured and insured consumers in high-deductible health plans.

The Price May Not Be Right: The Value of Comparison Shopping for Prescription Drugs

To assess physician views and perceived adoption of overtreatment guidelines and measure whether adoption of these guidelines influenced the recommendation of a targeted service.

A cross-sectional survey mailed from July 2014 to January 2015 to 902 internists who completed residency between 2003 and 2013, randomly selected from the American Medical Association Masterfile.

Poisson regression was used to model the rate of recommending a targeted service included in the guidelines, based on the level of guideline adoption. 

A total of 456 physicians responded (51% response rate). Most expressed familiarity with overtreatment guidelines (88.5%), a comfort level with discussing these guidelines with patients (79.9%), and described overtreatment guidelines as a useful tool in their practice (81.6%). Physicians in the highest tertile of guideline adoption reported double-digit rates of recommending antibiotics for sinusitis (29.7%), mammogram at end of life (16.5%), and electrocardiogram testing for asymptomatic patients (11.0%). Physicians in the bottom tertile of guideline adoption reported lower rates of recommending x rays (—12.0%; 95% confidence interval [CI], –19.4% to –4.5%; P = .002), magnetic resonance imaging for lower back pain (–4.8%; 95% CI, –8.1% to –1.5%; P = .004), and cardiac testing for asymptomatic patients (–10.2%; 95% CI, –18.9% to –1.5%; P = .02). 

US internal medicine physicians who completed residency between 2003 and 2013 reported high levels of adoption of overtreatment guidelines. Physicians who reported the highest levels of guideline adoption reported recommending services targeted by these guidelines in their practice.

We surveyed US internists who graduated residency in 2003 through 2013 about their views of overtreatment guidelines or recommendations against the use of potentially unnecessary tests and procedures. Key findings: 

US internists reported high levels of awareness, agreement, and use of overtreatment guidelines. 

Even physicians who reported the highest levels of guideline adoption, however, reported recommending services targeted by the guidelines in their practice. 

This research highlights the challenge of evidence-based de-implementation of medical tests and treatments in everyday practice.

Overtreatment in medicine, defined as “the waste that comes from subjecting patients to care that, according to sound science and the patients’ own preferences, cannot possibly help them,”

 is estimated to account for nearly 30% of healthcare spending.

 Increasing awareness that diagnostic and therapeutic interventions that physicians order are in some instances unnecessary

 has culminated in widely disseminated overtreatment guidelines, such as the Choosing Wisely campaign.

 Introduced in 2012, Choosing Wisely, an initiative of the American Board of Internal Medicine (ABIM) Foundation, has partnered with Consumer Reports and other medical organizations to provide physicians and patients with lists of potentially avoidable tests, treatments, and procedures.

In a 2014 telephone survey of 600 physicians, 38% of respondents reported having seen or heard about the Choosing Wisely campaign and 81% reported feeling “very comfortable” about “talking to patients about why they should avoid an unnecessary test or procedure.”

Despite the positive response from practicing physicians, there is little evidence that guidelines alone influence physicians’ ordering decisions. In fact, a recent report using commercial health plan claims data to evaluate the utilization of 7 services targeted by the guidelines failed to detect a meaningful decline in their use.

 However, the study looked at global use of services by health plan beneficiaries without accounting for physician characteristics. For example, a 2012 study of Massachusetts health plan data revealed that physicians with fewer than 10 years’ experience had the highest cost profiles compared with those of more senior physicians.

 Whether physicians’ awareness of overtreatment guidelines reduces the propensity to recommend a targeted service remains unknown.

To explore possible explanations behind the higher cost profiles of less experienced physicians, we surveyed recent internal medicine residency graduates about their adoption of overtreatment guidelines. Our specific objectives were to: 1) assess physician views of overtreatment guidelines using a novel 5-item scale, 2) estimate self-perceived practices according to select guidelines using hypothetical patient presentations, and 3) measure whether perceived adoption of the guidelines correlates with the likelihood to recommend a targeted service.

A literature review revealed no previously validated instruments evaluating physician attitudes toward overtreatment guidelines. To identify potential items for cognitive testing, we reviewed the literature, combed references from previously reported studies of physician views,

 and interviewed experts. Items were developed to assess: 1) physician awareness, agreement with, and use of overtreatment guidelines; 2) self-perceived propensity to recommend a service targeted by the guidelines; and 3) other potential confounders of physician practice identified in prior studies. We conducted 2 cycles of cognitive pilot testing to calibrate the wording to detect differences among physicians about these topics. After initial “think aloud” reviews with local (ie, Pennsylvania) practicing physicians, followed by revisions, we performed broader pilot testing with 100 internal medicine physicians randomly selected from the American Medical Association (AMA) Masterfile.

The final survey included questions about physician demographics, practice characteristics, attitudes known to influence overtreatment, views on overtreatment guidelines (awareness of, agreement with, and usefulness in practice), and self-reported practice in specific clinical scenarios related to the guidelines (eAppendix A [

]). Self-reported practice was assessed using fill-in-the-blank questions asking physicians to estimate the percentage of their patients they recommended for a specific test or treatment. Specifically, respondents were presented with brief descriptions of patient presentations for: 1) low back pain, 2) acute sinusitis, 3) cancer screening with a life expectancy of fewer than 10 years, 4) cardiac screening in asymptomatic routine care, and 5) a low pretest probability of venous thromboembolism (VTE). For example, when asked “For what percentage of patients with acute lower back pain do you order the following?” the respondent would fill in a percent for x-ray, magnetic resonance imaging (MRI), physical therapy, acetaminophen/anti-inflammatories, and opioids. The fill-in-the-blank questionnaire regarding treatment decisions has been shown to have high criterion validity (ie, it correlates with actual practice on similar patients) in prior studies.

To establish content validity, these items were tested by 11 clinical and survey design experts, including practicing primary care clinicians, researchers, and experts in survey design.

Using the AMA Masterfile, we pre-screened 2170 randomly selected internal medicine physicians who completed training within the last 10 years to confirm qualifying specialty, mailing address, and that the physician was actively seeing patients at least 20 hours a week. The final sample included 902 internal medicine physicians who were mailed a paper survey between July 2014 and January 2015 using a modified Dillman method.

 The initial mailing was done by first class mail accompanied by a $2 bill and followed by 2 reminder mailings approximately 6 weeks apart.

A set of 9 questions comprised the Overtreatment Guidelines Adoption (OGA) Scale, which assessed physicians’ attitudes toward overtreatment guidelines and cost containment in general. Six questions focused on awareness of, agreement with, and perceived usefulness of overtreatment guidelines; comfort denying patient requests for tests or treatments; comfort discussing costs with patients; and self-perception of cost consciousness. These were assessed using a 4-point Likert scale, ranging from strongly disagree to strongly agree. A second set of 3 questions assessed how frequently physicians discussed costs, used the overtreatment guidelines in practice, and found these guidelines to be useful. These were measured using a 5-point Likert scale of frequency.

To summarize overtreatment guidelines, their adoption, and measure physician attitudes toward guidelines separately from general attitudes toward cost containment, we developed 2 subscales: a 5-item OGA subscale and a 4-item cost-containment subscale, using standard factor analysis techniques (eAppendix B). The OGA subscale possible values ranged from 5 to 22; higher scale scores reflected a higher degree of adoption of overtreatment guidelines. The OGA subscale had high internal consistency with Cronbach alpha of 0.82 and rotated loadings of 0.44 to 0.75. Principal components analysis supported a separate cost-containment subscale of 4 questions related to costs (Cronbach alpha, 0.76; rotated loadings, 0.51-0.70).

Our main outcome measures were self-reported percentages of patients who were advised to elect 8 services targeted by 5 overtreatment guidelines. The guidelines were selected because they described common clinical scenarios in internal medicine, had been released at least 2 years prior to our survey, and were endorsed by multiple medical groups (

). We asked physicians to fill in the blank with the percentage of their patients they advised to receive a particular service. The options included services targeted by overtreatment guidelines, as well as other management options commonly offered to patients in each clinical context. The following tests and treatments were measured: x-ray and MRI imaging for acute low back pain; antibiotics for mild to moderate sinusitis; breast, prostate, and colon cancer screening for patients with a life expectancy of fewer than 10 years; electrocardiogram (EKG) testing for asymptomatic patients; and computerized tomography scan as the initial test for low-risk patients with possible VTE. Services that were recommended to less than 5% of patients (eg, Papanicolaou test for cervical cancer and stress test for cardiac testing in asymptomatic patients) were excluded from analysis.

Physician demographics, attitudes, reimbursement, and practice characteristics that may confound the relationship between physician views of, and practice according to, overtreatment guidelines, were included in the analysis. Physician demographics included age, gender, and race. Other physician characteristics included practice region, type of practice, compensation type, financial incentives (eg, quality, patient satisfaction, utilization review, productivity), insurance mix (eg, patients with Medicaid insurance), and attitudes (eg, comfort with clinical uncertainty, satisfaction with the practice of medicine in general, malpractice concerns). These items were either drawn from the AMA Masterfile (ie, age and gender) or included questions drawn from previously validated surveys of physicians.

Responses were entered in the REDCap electronic data capture tool (Harvard Catalyst; Boston), hosted at the University of Pennsylvania.

 Ten percent of entries were double-entered with perfect concordance. The data were exported in, and all analyses were conducted, using STATA version 13.0 (StataCorp; College Station, Texas).

We used the American Association for Public Opinion (Research Response Rate 2) definition.

Nonresponse bias was assessed by comparing respondents with nonrespondents and early to late respondents using the Pearson χ

The reported percentages of patients who were recommended for a particular test or treatment indicated a discrete number of events over a constrained range (0%-100%) and were positively skewed. Thus, the reported percentages were converted to a count variable based on a denominator of 100 (ie, 10% was converted to 10 out of 100) and modeled using Poisson regression. The independent variable of interest was a trichotomized OGA scale. Other variables in the model included a scale of physician attitudes toward cost containment in general (measured using the cost-containment subscale), physician demographics, practice characteristics, and attitudes previously shown to be associated with overuse (see “Other Variables” section of text). The predicted percentage of patients recommended for a particular test or treatment were estimated. Bootstrapping with 1000 iterations was used to estimate 95% confidence intervals (CIs). This study was reviewed and approved by the University of Pennsylvania Institutional Review Board.

Of the 902 potential respondents, 456 (51%) returned a completed survey. No differences between respondents and nonrespondents were observed by age, gender, region of current practice, or practice setting (

). Aside from Asian or Asian American respondents being overrepresented among late responders, there were no differences between early and late responders regarding gender, primary compensation, organization or setting of practice, or self-reported attitudes or satisfaction with medicine as a practice (eAppendix Table 1). Nearly half of the respondents self-characterized primary compensation type as salary with bonus (49.5%), followed by billings (28.1%) and salary only (20.9%), and the majority reported compensation linked to quality of care (62.9%) or productivity (65.1%) (

). Less than 5% of respondents (4.2%) completed residency in 2013. Other characteristics of the respondents’ practices are reported in Table 3.

Respondents’ attitudes toward cost containment are shown in 

. Most (88.5%) considered their practice style to be cost conscious. One in 4 (25.1%) reported discomfort discussing costs of care with patients, and 34.7% said they would not feel comfortable making a patient unhappy by denying a request for unnecessary care.

Respondents generally reported high levels of awareness, familiarity, and use of overtreatment guidelines (Table 4). Most (88.5%) reported being familiar with overtreatment guidelines in their specialty, 81.6% reported that the guidelines were useful in their practice, and 79.9% said they felt comfortable bringing up overtreatment guidelines in discussions with patients. However, less than 30% of respondents rated their agreement with these statements as “strong.” Respondents generally reported using overtreatment guidelines in practice with high frequency: 30.9% reported bringing up the guidelines in discussions with patients “frequently” or “always” and 44.2% reported bringing up the guidelines “occasionally.” Approximately 40% of respondents (41.1%) found the guidelines useful in practice “frequently” or “always,” and 42.4% found them “occasionally” useful. When individual responses were combined in the 5-item OGA subscale, the mean scale score was 15.6 (SD = 3.0) and the median 16 (interquartile range [IQR] = 14-18; observed range = 5-22).

In the fully adjusted models, respondents in the middle or top third of OGA subscale scores reported lower rates of recommending a test or treatment targeted by the guidelines for imaging for lower back pain, antibiotics for sinusitis, and cardiac testing for asymptomatic patients compared with the respondents in the bottom third of OGA scores (

). Physicians in the highest tertile of guideline adoption reported double-digit rates of recommending antibiotics for sinusitis (29.7%), mammogram at end of life (16.5%), and EKG testing for asymptomatic patients (11.0%). Physicians with OGA scores in the top third had significantly lower predicted rates of recommending x-rays (—12.0%; 95% CI, –19.4% to –4.5%; 

= .002) or MRI (—4.8%; 95% CI, –8.1% to –1.5%; 

= .004) for lower back pain and EKG for asymptomatic patients (—10.2%; 95% CI, –18.9% to –1.5%; 

= .02) compared with physicians in the bottom third of OGA scores. Physicians with OGA scores in the middle third also had lower predicted rates of recommending antibiotics for sinusitis (—6.9%; 95% CI, –13.0% to –0.8%; 

= 0 .03) and EKG for asymptomatic patients (—8.7%; 95% CI, –15.9% to –1.4%; 

= .02) compared with physicians in the bottom third of OGA scale scores. The differences in predicted probabilities across the tertiles of OGA scale scores were not significant for cancer screening and imaging as the initial test for patients at a low risk of VTE (Figure).

The association between physician cost consciousness and the percentage of patients recommended for a test or treatment targeted by the guidelines was not consistent: physicians in the top third of cost-consciousness scale scores reported lower rates of prescribing antibiotics for sinusitis and recommending mammography at the end of life, but this association was not observed for the other guidelines (eAppendix Table 2). Other factors associated with recommending services targeted by the guidelines were physician age, practice region, type and setting, practice that treated patients with Medicaid, and satisfaction with medicine as a profession.

In this survey study of physician views of overtreatment guidelines, internal medicine physicians generally reported high levels of awareness, agreement, and use of the guidelines in everyday practice, and their attitudes toward the guidelines were distinct from their attitudes toward cost containment. In addition, physicians who reported greater adoption of overtreatment guidelines recommended fewer tests or treatments targeted by some overtreatment guidelines, even after accounting for their overall cost consciousness. Physicians who reported the highest levels of guideline adoption, however, also reported recommending services targeted by the guidelines in their practice.

In a national survey, US internists reported high levels of adoption of overtreatment guidelines; despite this fact, they also reported recommending services targeted by the overtreatment guidelines.

July 2017

The Association Between Insurance Type and Cost-Related Delay in Care: A Survey

The Price May Not Be Right: The Value of Comparison Shopping for Prescription Drugs

US Internists' Awareness and Use of Overtreatment Guidelines: A National Survey

Availability and Variation of Publicly Reported Prescription Drug Prices

Cost-Effectiveness of a Patient Navigation Program to Improve Cervical Cancer Screening

Twitter Accounts Followed by Congressional Health Staff

Predicting High-Cost Privately Insured Patients Based on Self-Reported Health and Utilization Data

Adaptation of an Asthma Management Program to a Small Clinic

Medication Adherence and Improved Outcomes Among Patients With Type 2 Diabetes

Leveraging EHRs for Patient Engagement: Perspectives on Tailored Program Outreach