a b s t r a c t

Objective: The purpose of this study is to determine the economic value of the Utah Poison Control Center (UPCC) by examining its contribution to the reduction of unnecessary emergency department (ED) visits and associated charges across multiple years.

Methods: A multi-year (2009-2014) analysis of cross-sectional data was performed. Callers were asked what they would do for a poison emergency if the UPCC was not available. Healthcare charges for ED visits averted were calculated according to insurance status using charges obtained from a statewide database.

Results: Of the 10,656 survey attempts, 5018 were completed. Over 30,000 cases were managed on-site each year. Using the proportion of callers who noted they would call 911, visit an ED, or call a physician’s office, between

20.0 and 24.2 thousand ED visits were potentially prevented each year of the survey. Between $16.6 and $24.4 million dollars in unnecessary healthcare charges were potentially averted annually.

Conclusions: Compared to the cost of operation, the service UPCC provides demonstrates economic value by re- ducing ED visits and associated charges. As the majority of patients have private insurance, the largest benefit falls to private payers.

(C) 2016

Introduction

Poison control centers (PCC) across the US bring value to the com- munities they serve [1]. The clinical value provided by PCCs, such as the reduction in hospital length of stay, guiding care given during emer- gency department (ED) visits, and preventing unnecessary healthcare utilization, contribute to the economic value of PCCs [1]. Previous re- search demonstrates that PCCs are on par with the cost savings generat- ed from pediatric immunizations, saving $6.50 for every $1 spent [2].

Multiple studies have quantified the cost-savings generated by PCCs across different regions in the US [3-7]. To date, however, much of the research assessing value has utilized cross-sectional data. To understand more fully the value that PCCs may contribute economically, it is impor- tant to examine data spanning more than one year. This will help ac- count for fluctuations in the data that may occur naturally, such as volume of cases, volume of ED visits, and costs. The purpose of this study is to determine the economic value of the Utah Poison Control Center (UPCC) from the perspective of payers by examining its

* Corresponding author at: 30 S 2000 E, Rm 4410, Salt Lake City, UT 84112, United States.

E-mail address: [email protected] (C.R. Tak).

contribution to the reduction of unnecessary ED visits and charges across multiple years.

Methods

The Utah Poison Control Center is an AAPCC-accredited center that covers the state of Utah. The UPCC responds to more than 40,000 human poisoning exposure inquiries each year, providing health educa- tion, medical guidance, and assisting care providers with treatments.

Survey

Since 2002, UPCC has conducted ongoing customer satisfaction sur- veys to ensure that it continues to meet the needs of the community. Pharmacy students employed at the UPCC conduct surveys with the in- dividual reporting the poison exposure case. A poison exposure is de- fined as any contact with a potential toxic substance. Surveys are only conducted when the poison exposure case is managed on-site in a non-Healthcare facility. Approximately 6% of cases of all ages are ran- domly chosen by computer to survey. Every attempt is made to conduct surveys within 2 weeks of the initial contact with the poison control center. An analysis of surveys conducted from January 2011 through

http://dx.doi.org/10.1016/j.ajem.2016.11.049

0735-6757/(C) 2016

September 2014 found a median of 8 (IQR 6-10) days from initial con- tact to survey.

For this survey, the surveyor (pharmacy student) reads the 26 ques- tions to the caller. Most questions are close-ended (e.g. yes/no or Likert scale); however some question are open-ended, which are coded by the surveyor into pre-defined categories. Space is also allowed for free re- sponse. Among the questions asked, an open-ended question is dedicat- ed to identify what, if any, action would have been pursued had the UPCC not been available. The original survey included items 1-3 and 10 (see below). In 2014, items 4-9 and 11 were added based on previ- ous verbatim responses. Responses from the survey are aggregated quarterly. Categories for responses are as follows:

Call 911

Call physician

Visit ER

Visit urgent care

Call pharmacist

Call nurse hotline

Call family/friend

Read label

Search online

Other

Don’t know

The survey also collects insurance status of the patient through a close-ended question. Options include Medicaid, Medicare, Children’s Health Insurance Program (CHIP), Other State Insurance, Private Insur- ance, Federal Insurance, Other Insurance, No Health Insurance, Refused, and Unknown. In our analysis, we grouped insurance status into five categories, Medicaid, Medicare, Private, no insurance, and Other Insur- ance (including CHIP, Other State Insurance, Federal Insurance, Other Insurance, Refused, and Unknown).

We conducted our analysis using the results from the 2009-2014 surveys. The quarterly results were aggregated, weighted according to number of respondents, and an average annual response rate was calcu- lated. To obtain the proportion of persons with an exposure who would utilize ED services if the UPCC were not available, we combined the pro- portion of callers who responded that they would have made a “visit to the ED” with the proportion of callers who indicated that they would have called 911 and then been transported by EMS . This is supported by previous research that suggested the ma- jority of persons who summon EMS for an exposure are taken to an ED [3].

To ensure that we were only including non-healthcare provider cal- lers with an exposure that was managed on-site, an internal report was used to identify human exposure management site and excluded healthcare facility (MD office, acute care, free standing, urgent care and other [group homes]). We obtained the proportion of callers in each year that were managed on-site. For this analysis, it was important to exclude callers who would have needed further medical services as we were trying to examine the value of the UPCC in reducing unneces- sary healthcare utilization.

To calculate the number of potential cases prevented by the UPCC, the total number of cases with an exposure managed on-site was mul- tiplied by the proportion of survey callers that indicated they would have used ED and EMS services.

Financial data

The financial data used for this analysis were obtained from the Indi- cator-Based Information System for Public Health (IBIS-PH), a statewide public database that collects data on disease prevalence, disease inci- dence, injuries, and hospital charges [8,9]. Queried data are stratified by geographic location, demographic information (e.g. age, gender, in- surance status.) and year, among other potential variables [9]. The most recent year of hospital charge data available was 2014. We

obtained hospital charges related to unintentional poisoning injuries by year and insurance status (Medicaid, Medicare, Private, No Insur- ance, and Other) for uncomplicated cases treated and released from an emergency department. Hospital charges represent an aggregate of charges originally billed for specific diseases/conditions on all hospital claims in the state and include a median, a mean, and a range. The me- dian value was chosen for this analysis. For the Private Insurance cate- gory, the median charge for the largest private payer by volume in the database was used. For the Other Insurance category, the aggregated median charge of all payers was used. All charges were adjusted to 2014 US Dollars (USD) using the personal health care expenditures (PHCE) index [10]. Individual cost data were not available for this analysis.

To obtain the potential charges averted, the median hospital charges were multiplied by the number of UPCC cases who would have visited the ED had the UPCC not been available. This was done for each year in the analysis and stratified according to insurance status. We were able to use aggregated insurance status of patients to get a proportion of each Type of insurance. This enabled us to calculate a more accurate representation of the potential charges averted. Insurance status was available starting in 2010. For 2009, we used the proportion for each in- surance status from callers from 2010.

To provide a comparison of the relative value of charges averted to operational cost, the 2011 average expenses of poison control centers in the US were inflated to 2014 USD using the Gross domestic product (GDP) deflator [1,10]. The University of Utah Institutional Review Board approved this study.

Other analyses

Our base case examined patients that would have sought care through ED and/or EMS services. In a separate analysis, we included the proportion of patients who indicated they would call their physi- cian. A study by Kearney et al. demonstrated that physicians’ offices would recommend “go directly to ED” or “call 911” 64% of the time if PCC services were not available [5].

Results

We had 10,656 survey attempts, of which 5018 (47.1%) callers com- pleted the survey over the years of 2009 to 2014. The callers and their responses are summarized in Table 1.

For each year, between 30.9 and 36.9 thousand non-healthcare cal- lers were managed at home (Fig. 1). Using the proportion of those cal- lers who would have visited an ED had the UPCC not been available, we calculated that each year between 14.4 and 16.8 thousand potential ED visits were prevented. Including the proportion of patients who

Table 1

survey responses of PCC callers of home-managed cases by insurance status and year (2009-2014).

	2009	2010	2011	2012	2013	2014
Total surveyed	2333	1534	1728	1653	1504	1904
Completed surveys	1080	751	831	856	660	840
% Completed	46.29	48.96	48.09	51.78	43.88	44.12
Insurance status (%) Medicaid	15.98a	15.98	16.95	16.08	13.48	12.80
Medicare	5.46a	5.46	5.77	7.27	5.60	5.46
Private	58.59a	58.59	58.29	56.48	62.58	67.98
No insurance	9.19a	9.19	9.98	7.54	7.73	4.76
Otherb	10.78a	10.78	9.01	12.63	10.61	9.01
Responses (%)
Call 911	24.39	21.14	19.96	24.10	19.99	14.74
Visit ER	21.11	27.92	30.65	28.51	26.93	29.52
Call Physician	31.46	26.84	26.48	28.08	28.14	32.11

^a 2010 insurance status.

^b Represents the aggregated median charge for all payers from IBIS-PH database.

10000

5000

0

16797

16963

16521

16573

15000

14503

14114

20394

2009

2010

2011

2012

2013

2014

Homecare-Managed Human Exposures Potential ED Visits

Potential ED Visits Including those who would have called their physician

20000

20001

22258

22077

22909

24216

30000

25000

31890

30913

31504

32644

34580

35000

36921

40000

Number of Human Exposure Cases and ED Visits by Year (2009-2014)

Fig. 1. Number of human exposure cases and ED visits by year (2009-2014).

would contact a physician, we calculated that each year between 20.0 and 24.2 thousand ED visits were prevented.

Hospital charges relating to unintentional poisoning injury ranged from $759 (Medicaid 2009) to $2368 (Medicare 2011), adjusted for in- flation using the PHCE. The charges separated by payer type are summa- rized in Fig. 2.

In Fig. 3, we show the potential charges prevented according to year and insurance status. In the base case, from $16.6 to $24.4 million in charges were potentially averted. This is compared to an average annual expense of approximately $2.5 million for a certified poison center [1]. As the majority of the callers had private insurance, this is where most of the charges fall. Total potential charges prevented are also reported. When callers who would call their physician are included in the analysis, the proportion of patients visiting the ED has an absolute in- crease of 0.17-0.20 (e.g. from 0.46 to 0.66 - see Fig. 1 for the change in potential ED visits); therefore, likely estimates of cases prevented are even higher with still greater charges prevented. See Fig. 4 for

these results.

Discussion

We were able to demonstrate that from 2009 to 2014, the UPCC po- tentially averted $16.6-$24.4 million dollars annually in unneeded medical charges to government and private payers and self-pay pa- tients. When compared to the average annual expenses of a certified poison center, the service UPCC provides results in profound value for payers. These findings can aid decision-makers when determining allo- cation of resources to PCC services.

The results from the survey reveal some interesting trends. The number of cases to the UPCC shows a downward trend. In 2014, the UPCC was consulted on the management of approximately 41,000 poi- sonings cases, down from the 47,000 cases in 2009. This could be due to a variety of factors, such as a decline in birth rate. The number of births per 1000 population in Utah has declined from 21.1 in 2004 to 17.4 in 2014 [8]. Since 60% of calls to the UPCC involve children younger than 6 years old, the Number of calls to the UPCC would follow that decreas- ing trend. Additionally, it could be that people are searching the internet

Median Hospital Charges in 2014 USD

by Insurance Status and Year (2009 to 2014)

$2,500.00

$2,000.00

$1,500.00

$1,000.00

$500.00

$0.00

2009

2010

2011

2012

2013

2014

Medicaid Medicare Private No Insurance Other

Fig. 2. Median hospital charges in 2014 USD by insurance status and year (2009 to 2014).

Medicaid Medicare Private No Insurance Other

2014 Certified Poison Center

Average Annual Expenses

2013

2012

2011

2010

2009*

$30,000,000

$25,000,000

$20,000,000

$15,000,000

$10,000,000

$5,000,000

$0

Base Case Analysis of Total charges Prevented by Insurance Status and Year (2009-2014)

Fig. 3. Base case analysis of total charges prevented by insurance status and year (2009-2014).

more often for information. Our survey results show that in 2014, 14.9% of respondents chose “search the internet” if the UPCC didn’t exist. Even with the reduction in cases, the UPCC consults on an average of 14 cases per 1000 population, which is twice the national average for poison centers.

Despite the decrease in cases per year, the proportion of callers who would have used the ED remained approximately the same each year, with some fluctuation but no apparent trend. However, there was a de- cline in responders who indicated they would call 911. It is not clear if this represents a true decrease in the number of people that would uti- lize Healthcare resources (call 911 or go directly to the ED) or represents increase use of non-healthcare resources such as searching the internet. During the period of our study, there was a statewide media campaign

to reduce inappropriate use of 911 resources. Even more striking, we found that although 2013 had the fewest callers of any of the years, it did not have the overall lowest hospital charges. This appears to be due, at least in part, to the rise in hospital charges that was seen over the years.

The percentage of UPCC patients covered by Medicaid in our study period is 2.2% to 5.6% above the state estimates but 0.6% to 6.4% below national estimates [11]. State and national estimates of uninsured were similar between 2010 and 2014 [11]. The percentage of UPCC pa- tients who are uninsured is 5.3% to 7.7% below the state estimates and 5.1% to 7.3% below national estimates [11]. This suggests that dispropor- tionately fewer persons who are uninsured in Utah are utilizing the UPCC as compared to the insured population and may result in an

Analysis with Proportion of physician referral of Total Charges Prevented by

Insurance Status and Year (2009-2014)

$40,000,000

$35,000,000

$30,000,000

$25,000,000

$20,000,000

$15,000,000

$10,000,000

$5,000,000

$0

2009*

2010

2011

2012

2013

2014 Certified Poison Center

Average Annual Expenses

Medicaid Medicare Private No Insurance Other

Fig. 4. Analysis with proportion of physician referral of total charges prevented by insurance status and year (2009-2014).

unnecessary Financial burden on these patients. Efforts are needed to ensure that all residents, regardless of insurance status, are aware of poison center services.

There are some limitations in our study. Although our data are accu- rate with regard to the caller responses, we had no way of knowing if the callers’ intentions would have been realized. As the majority of sur- veys were conducted approximately one week after the case was re- solved, we were not able to capture the response of the caller as it related to their perceived severity of the situation at the time of the call. This may have influenced some callers to choose an alternative ac- tion as a survey response that may have been more attenuated given their full insight into the poisoning episode - which was treated as “homecare.” This is in comparison to the limited insight that callers have at the time of the call, which may be a more severe perception of the situation. Therefore, we may be underestimating the proportion of patients who would have chosen to seek EMS. Additional research is needed to examine the accuracy of these responses and the effect that it may have on the monetary savings generated by a PCC.

Additionally, there were no individual-level data available for the analysis. Therefore, we could not assess any differences in the propor- tion of callers who would have sought medical services based on insur- ance status. It is unknown to us if patients with different insurance statuses (i.e. private vs public) have the proclivity to choose alternative scenarios in a differential fashion. We also were not able to examine dif- ferences between pediatric and adult patients. Although we weren’t able to determine the generalizability of our sample to the UPCC patient population as a whole, the patients surveyed were chosen at random. This maximizes the generalizability of our results to the larger UPCC population.

For the analysis, we decided that all patients who call EMS should be grouped with patients who declared that the ED was the option of choice if the PCC hadn’t been available. This is supported by LoVecchio et al. who found that nearly all persons who call EMS are transported to the ED [3]. However, despite including this proportion of patients, we did not include the associated EMS charges. The transport charges in Utah range from $696 to $1344, plus mileage [12]. This renders our estimates conservative in charges averted. We also excluded the pro- portion of patients who would call their medical provider in the ED group in the base case. We excluded this group because we did not col- lect data on what a provider would do if faced with a patient who had a potential poison exposure. This is a limitation in that it does not reflect as accurately the reality of patient care and grossly underestimates the value of a PCC. We did include it, however, as part of separate analysis (Fig. 4). Some authors report that the majority of medical care providers in primary care offices would instruct their patients to go straight to the ED or call 911, regardless of the type of poisoning [5,13]. We reported our results based on the results from Kearney et al. [5]; however, Austin et al. found that the proportion was higher at 82.5% [13].

The financial data available to us for this analysis included hospital charges from IBIS-PH and does not include provider charges. These charges do not reflect the amount paid by the patient or the patient’s medical insurance (both public and private). Charges are typically greater than the amount paid. According to a study by Hsia et al., reim- bursement rates in 2003 for non-hospitalized pediatric patients range from 35% for Medicaid to 58% for private insurance [14]. Henneman et al. found that revenue for Medicaid and private insurance ranges from

$111 to $273 and from $102 to $1281, respectively, depending on the fa- cility billing level [15]. Charges, therefore, inflate our results and give us larger sums averted than what would likely be the case with Actual costs. Taking into account reimbursement rates as outlined by Hsia et al., the UPCC averted $8.2 to $13.1 million in potential unnecessary re- imbursement from payers.

Additionally, this estimate does not take into account the economic value of the UPCC in reducing hospital length of stay as we only exam- ined cases that were treated and released. In a study by Vassilev and Marcus, patients hospitalized for a poison-related injury who received

PCC assistance had a mean (median) reduced hospital stay of approxi- mately 3.02 (3.0) days [16]. The authors report that the average per day charges of a hospital stay in New Jersey are approximately $6000 [16]. Reducing a hospital stay by 3 days could prevent $18,000 in charges. This would only reinforce the cost-saving conclusions of our analysis.

Furthermore, for patients who are referred by the PCC to the ED, this analysis does not estimate the economic value of the poison center in identifying a patient with a poisoning exposure who gets immediate and prompt attention, and averts the additional costs associated with a Delayed presentation of a serious poisoning.

Indirect costs such as loss of work time and caregiving were not in- cluded in the study, but would likely increase the overall savings due to PCC utilization.

There are also a number of strengths in this study. We used survey results collected within approximately 8 days of contact with the UPCC to assess the hypothetical behavior of callers of the PCC. This is im- portant as it captures more accurately what individuals would do at the time of the event. This likely minimizes any recall bias that callers may have and therefore gives more accurate estimates. Additionally, multi- ple years of data were available for our analysis. This allows us to get a long-term perspective of the value of a PCC as opposed to a single- year, cross-sectional view.

We found that the UPCC provides a substantial financial benefit to private and public payers in the form of reducing unnecessary ED visits across multiple years. Additional research is needed to look at the actual cost-savings over multiple years using real-world cost data that can ex- amine the nuances of insurance status and pediatric vs adult utilization. Further, this study only examined the preventive nature of the PCC for callers with a potential poison exposure. What is unknown is how much more value a PCC may offer for persons who do not contact the PCC but instead unnecessarily go straight to EMS or a local ED. Addition- al research is needed to identify the encounters that could be prevented by PCC involvement.

Sources of funding

No source of funding was obtained for this project.

Conflict of interest

The authors report no conflict of interest.

Previous presentation of data

None.

Acknowledgments

The authors thank B. Zane Horowitz, MD, interim medical director of the Utah Poison Control Center, and Joseph Biskupiak, PhD, for reviewing this manuscript.

References

1. The Lewin Group I. Final report on the value of the poison center system. American Association of Poison Control Centers; 2012.
2. Miller TR, Lestina DC. Costs of poisoning in the United States and savings from poi- son control centers: a benefit-cost analysis. Ann Emerg Med 1997;29(2):239-45.
3. LoVecchio F, Curry S, Waszolek K, Klemens J, Hovseth K, Glogan D. Poison control

   centers decrease emergency healthcare utilization costs. J Med Toxicol 2008;4(4): 221-4.

   Phillips KA, Homan RK, Luft HS, et al. Willingness to pay for poison control centers. J Health Econ 1997;16(3):343-57.

4. Kearney TE, Olson KR, Bero LA, Heard SE, Blanc PD. Health care cost effects of public use of a regional poison control center. West J Med 1995;162(6):499-504.
5. Blizzard JC, Michels JE, Richardson WH, Reeder CE, Schulz RM, Holstege CP. Cost- benefit analysis of a regional poison center. Clin Toxicol (Phila) 2008;46(5):450-6.
6. Zaloshnja E, Miller T, Jones P, et al. The impact of poison control centers on poison- ing-related visits to EDs-United States, 2003. Am J Emerg Med 2008;26(3):310-5.
7. Utah Department of Health CfHDaI. Public Health Indicator Based Information Sys- tem (IBIS). https://ibis.health.utah.gov/; 2016.
8. Braithwaite G, Haggard LM. Utah’s IBIS-PH: an innovative user interface solution for Web-based data query systems. J Public Health Manag Pract 2006;12(2):146-54.
9. Quality AfHRa. Using appropriate price indices for analyses of health care expendi- tures or income across multiple years. Medical Expenditure Panel Survey. https:// meps.ahrq.gov/about_meps/Price_Index.shtml; 2016. [Accessed October 12, 2016].
10. Barnett JC, Vornovitsky MS. Health insurance coverage in the United States: 2015. Washington, DC: U.S. Government Printing Office; 2016.
11. Bureau of Emergency Medical Services UDoH. Ambulance rates. Utah EMS law and

    administrative rules 2015; https://health.utah.gov/ems/rulereg/. Accessed March 24, 2016, 2016.

    Austin T, Brooks DE, Welch S, Lovecchio F. A survey of primary care offices: triage of poisoning calls without a poison control center. International Journal of Family Med- icine 2012;2012:417823.

12. Hsia RY, MacIsaac D, Palm E, Baker LC. Trends in charges and payments for nonhos- pitalized emergency department pediatric visits, 1996-2003. Acad Emerg Med 2008;15(4):347-54.
13. Henneman PL, Nathanson BH, Li H, et al. Is outpatient emergency department care profitable? Hourly Contribution margins by insurance for patients discharged from an emergency department. Ann Emerg Med 2014;63(4):404-11 (e401).
14. Vassilev ZP, Marcus SM. The impact of a poison control center on the length of hos- pital stay for patients with poisoning. J Toxicol Environ Health A 2007;70(2): 107-10.