Advice for lay callers with low-risk poison exposures by a regional poison control center: the impact on health care expenditures

Decision analysis

A decision tree model consistent with the simplified treatment pathways for exposures of laypersons (i.e., general public, no institutional callers) with poisoning concern was constructed (Fig. 1). The objective was to compare the costs of operating the PCC with its monetary benefit. For this, the perspective of the German health care system, including public health expenditures as well as the costs of private and statutory health insurance companies was taken. Analogous to comparable studies, the benefit was defined as the costs that can be avoided through the existence of the PCC. In the decision analysis, a situation utilizing PCC consultation and without PCC consultation was compared by calculating the average weighted financial outcome of each scenario. All input parameters used for the analysis can be found in the tables of Additional files 1 and 2. Calculations were performed using Microsoft Excel 2016 (Microsoft Corporation, Redmond, Washington, USA). The conducted analysis was based on the following assumptions:

(1)

All persons who call a medical doctor cause costs for medical advice. If the case is impossible to be solved by telephone, the medical doctor will send the affected person to the closest emergency department. Callers are not asked to come to the doctor's office for treatment.

(2)

If a layperson has already called the PCC, the attending medical doctor will not call the PCC again for this case.

(3)

Transition probabilities that could not be derived from the survey are assumed to be identical for the scenario with and without PCC consultation.

Fig. 1figure 1

Decision tree with base case input values for low-risk poison exposures of laypersons in a situation utilizing PCC consultation and a hypothetical situation without PCC consultation. *Value derived from survey results; **Value derived from Berlin poison control center's data (2019); ED Emergency Department, EMS Emergency Medical Services, MD Medical Doctor

Determination of transition probabilities

The data for calculating transition probabilities were collected through a survey of former callers at the Berlin PCC (Charité, Berlin). The Berlin PCC is handling about 45,000 human poison exposure cases per year. About half of these inquiries are made by laypersons. Most lay inquiries concern children aged ≤ 12 years (approximately 78%). An age definition ≤ 12 years for children is also used by the "Einheitlicher Bewertungsmaßstab (EBM)", a nationwide catalog on which the reimbursement of services provided by “National Association of Statutory Health Insurance Physicians (NASHIP)” accredited doctors in Germany is based on. Since most lay calls involve low-risk poison exposures, home management (i.e., self-monitoring at home) can be recommended in approximately 90% of the cases. The remaining cases are moderate to severe poisonings that require urgent medical help.

Interview of Berlin poison control center's callers

Since the COVID-19 pandemic reached Germany in early 2020, there has been a distinct change in the use of public health facilities [20]. Due to this fact, only data from before the COVID-19 outbreak in Germany were used for the study. The sample was limited to calls between December 2019 and January 2020 to ensure that respondents would remember their last call to the Berlin PCC. For the survey, only calls from laypersons regarding poisoning exposures of humans were included in which management at home was recommended, as this group reflects most of the laypersons calling the PCC.

For ethical reasons, cases with suicidal or criminal intent were not considered. Consequently, 2,419 cases were potentially eligible for the survey. Using the “KNIME Analytics Platform 4.2.2” (KNIME AG, Zurich, Switzerland) [21] 1,500 cases were randomly selected. A standardized questionnaire for the telephone survey was designed by the authors, which was pretested with 30 persons for content and comprehensibility by one of the authors. The same person, who performed the pretest, conducted a retrospective telephone survey for a five-week period between November 9 and December 15, 2020 subsequently. The aim was to acquire a minimum of 377 survey respondents, including the 30 pretested persons, to obtain a sufficiently large sample size. The sample size was calculated based on a total population of 19,378 calls in 2019 that met the criteria described previously (confidence level: 95%, margin of error: 5%). All potential survey participants were called up to three times at different times of the day.

After explaining the purpose of the survey and obtaining informed consent, the participants were asked: 1) Which medical service/s he/she had used before or after calling the PCC, 2) What he/she would have done if the PCC had not been available, 3) Whether the affected person was covered by private or statutory health insurance at the time of the call.

If a survey participant could not answer one of the open questions, the possible response options were read to the interviewed person. All responses were recorded in a standardized manner so that a systematic evaluation was possible after completing the survey. All persons who did not consent or were unable to participate (e.g., could not remember the former PCC call) were excluded from the survey.

During the evaluation, the response options "Call physician," "Call hospital," and "Call 116,117Footnote 1" were merged in the main category "Call a medical doctor (MD)" to simplify the analysis. Moreover, the response options "Do research on the internet," "Call family/friends," "Do nothing/other," "Call a pharmacy," "Read the instructions on the packaging," and "I do not know" were combined in the main category "Do nothing/other" (Fig. 1 and Additional file 1).

Analysis of Berlin poison control center's data

Using the PCC's data (2019), the probabilities for the branches “Managed by MD/MD sends to ED”, “Outpatient treatment/Inpatient treatment” and “Managed by emergency service/Emergency service sends to ED” were determined, as they could not be obtained from the survey. To calculate the probabilities, the procedure recommended by the Berlin PCC for the different caller categories (doctors' office, hospital staff, and emergency services) were analyzed. Due to the fact that most lay inquiries have a low risk and concern children, only cases with asymptomatic patients aged ≤ 12 years were included to calculate transition probabilities (Fig. 1 and Additional file 1).

Determination of costs and charges

All costs and charges used in the analysis are stated in Euro.

The variable and fixed costs per case were taken into account to calculate the PPC's average cost for one lay consultation. Based on a mixed calculation, staff costs of €14.79 were determined to process and follow-up a layperson's inquiry. Overhead costs (e.g., administrative and IT staff, rate for space rental and utilities, technical equipment, office supplies) amount to €21.40. Overall, this results in total costs of €36.19 per case.

By using the EBM catalog, charges for the treatment of persons with statutory health insurance were identified. Costs for privately insured persons were considered by multiplying corresponding EBM charges by a factor of 2.28 [23]. According to our own empirical data, most requests address children. For that reason, charges for the treatment of children (≤ 12 years) rather than for adults were used whenever possible. To determine appropriate charges according to the EBM catalog a pediatrician, a general practitioner, and a head of an emergency department for children was interviewed. Except for the branches "Call MD → Managed by MD" and "Call MD → MD sends to ED," no surcharges were included for the use of medical services outside of consultation hours and on weekends. Moreover, only direct costs were considered and indirect costs were excluded (e.g., driving and waiting time, increased length of stay in the hospital, work loss days). Charges for the actions “Call 112 (EMS)Footnote 2” and “Call 112 (EMS) → Managed by emergency service/Emergency service sends to ED” were calculated based on data of the Berlin Fire Department [24]. The average costs for “Outpatient treatment” were derived from the literature [25], while the average costs for “Inpatient treatment” could be determined based on Charité controlling department data (2019). The actions “Management at home,” “Go to MD,” “Call MD,” “Go to hospital” and “Do nothing/other” were considered to be free of charge. Additional file 2 gives a detailed overview of all costs and charges included in the analysis.

Sensitivity analysis

A sensitivity analysis based on the input parameters presented in the tables of Additional files 1 and 2 was conducted to test the robustness of study results to changes in cost drivers and probabilities. For that, IBM SPSS Statistics 27 (IBM Corporation, Armonk, New York, USA) and Microsoft Excel 2016 (Microsoft Corporation, Redmond, Washington, USA) was used. For cost drivers, a variation of  ± 25% was assumed. Based on the survey results and PCC's data (2019), the minimum and maximum values (= upper and lower bound of 95% confidence interval) for the transition probabilities were calculated using the bias-corrected and accelerated (BCa) bootstrap method with 1,000 re-samples. A deterministic one-way sensitivity analysis (DSA) for each input parameter (base case and min./max. values) with consideration to the recommended procedure of Sendi and Clemen [26] for chance nodes with more than two branches was performed. To be able to take 0%-probabilities into account, a min./max. range of 0% to 1% was estimated for them in the DSA. In addition to the DSA, a probabilistic sensitivity analysis (PSA) was conducted using a Monte Carlo simulation with 1,000 iterations and a normal distribution assumption.

留言 (0)

沒有登入
gif