Background: The implantation of an Medication Rec- onciliation (MR) process has been shown to significantly reduce ME in the adult population. Although MR is an in- creasingly consolidated activity in adults, the same cannot be said of the pediatric population, which is a group particu- larly at risk of suffering ME.

      Aim of the review: To implant a MR model in pediatric patients as a high-risk population, in order to reduce Medi- cation Errors (MEs) upon hospital admission.

      Method: The project was carried out in a pediatric hospi- tal between January-November 2018 by a pharmacist spe- cialized in Hospital Pharmacy. A daily list was compiled of the patients admitted in the last 24 hours, and in each case the Best Possible Medication History (BPMH) was obtained for comparison against the treatments actually prescribed upon admission, with the aim of analyzing possible discrep- ancies. Pharmaceutical Interventions (PIs) were moreover carried out to avoid MEs.

      Results: A total of 1760 patients (45% females and 55% males, mean age 7.9 ± 5.2 years; range 12 days-24 years) were reconciled. Sixty percent presented background dis- ease. Fourteen percent received more than four drugs in home treatment (range 4-15), including some drug with a Narrow Therapeutic Index (NTI). A total of 830 discrep- ancies were detected in the prescriptions of 592 patients (34%). Of these, 307 (37%) were justified by the clinical condition of the patient, while 523 (63%) were not justified and were classified as Reconciliation Error (RE). These REs were detected in 334 patients (56%). The main drug groups involved were psycholeptics and psychoanaleptics (n= 58), antiepileptic drugs (n=57) and systemic antibacterials (n=51). A total of 460 PIs were made, of which 72% received immediate acceptance.

      Conclusion: Medication reconciliation in pediatrics is able to detect a significant percentage of errors, preventing them from reaching the patient.

      Impact on Practice: Our study shows that MR in pedi- atrics, in the same way as in adult patients, is useful for the detection of MEs, preventing them from reaching the pa- tient. Fifty-six percent of peadiatric patients has at least one reconciliation error. The main drug groups involved were psycholeptics and psychoanaleptics.

      Patient safety is a priority concern and a key aspect of health- care. The use of drugs is a complex process in which Medication Errors (MEs) may occur. These are defined as any avoidable in- cident that can cause harm to the patient or give rise to inap- propriate drug use, when medication is under the control of the medical professional, the patient or consumer. Such MEs have a great impact upon patients in terms of morbidity-mortality and occur particularly during patients transition from one health- care level to another. The ENEAS study [1] (National Study of Adverse Events), published in 2005 in Spain, showed 37.4% of the adverse events in hospitalized patients to be related to the prescribed medication. Likewise, the EVADUR study, conducted in the emergency care setting, found the second most frequent cause of adverse events to be medication practice [2].

      For these reasons, world health institutions such as the Joint Commission on Accreditation of Healthcare Organizations (JCA- HO) and the World Health Organization (WHO) consider Med- ication Reconciliation (MR) to be one of the solutions to this problem. In this regard, MR is defined as the formal and stan- dardized process of obtaining the complete list of the previous medications of patients for comparison against active prescrip- tion, with analysis and solution of the observed discrepancies in order to guarantee that the patients receive all the drugs need- ed for chronic treatment adapted to their current clinical condi- tion and to the new prescriptions made upon hospital admis- sion. Accordingly, the objective of MR is to detect and prevent ME, thereby increasing patient safety and the effectiveness of treatment [3,4]. According to the 2020 Initiative of the Spanish Society of Hospital Pharmacy (SEFH), “Towards the future, with safety”, there should be standardized MR procedures upon ad- mission and at discharge in all hospitals [5].

      Different guides for the implantation of MR have been pub- lished in Spain, with recommendations on the need to incor- porate a standardized system for compiling the information on all the patient drugs or contemplate the use of technological resources to facilitate the reconciliation procedure [6,7]. The Emergency Pharmaceutical Care working group (REDFASTER) found that 79% of the patients in the emergency care depart- ment presented discrepancies between the medication they were actually receiving before admission to the hospital emer- gency care department and the medication reflected in their case histories [8]. Furthermore, the abrupt suspension of cer- tain drugs may result in withdrawal syndrome or exacerbation of the background disease, thereby complicating the acute con- dition leading to admission to emergency care, or even generat- ing a new health problem. The criteria of the REDFASTER for a patient in the emergency care department to be regarded as a candidate for reconciliation are [9]:

• Admission to emergency care for over 24 hours, with the existence of routine home

      Estimated admission to emergency care for less than 24 hours but with the existence of routine treatment involving drugs that should be reconciled in a period of under four hours (Table 1).

      The implantation of an MR process has been shown to signif- icantly reduce ME in the adult population, and in this regard the hospital pharmacist is the professional best suited for carrying out the process [10,11]. Different authors consider that given the knowledge and experience of hospital pharmacists, these professionals should be in charge of leading MR [12], and can contribute to improve the clinical and economic outcomes [13- 17]. Other authors affirm that in view of the evidence on the benefits of MR in patients, reconciliation should be established as a priority activity on the part of hospital pharmacists [18].

      Cornish et al., analyzed prescription practice in 151 adult pa- tients admitted to Internal Medicine with at least four drugs as routine home treatment. They identified at least one unintend- ed discrepancy in 53.6% of the patients – the most prevalent being the omission of some drug [19].

      Although MR is an increasingly consolidated activity in adults, the same cannot be said of the pediatric population, which is a group particularly at risk of suffering ME. Furthermore, the like- liness that such errors will cause adverse events is up to three times greater than in the adult population [20-22]. This is due to the existence of risk factors inherent to the pediatric popu- lation: significant differences in body composition and physiol- ogy in children versus adults; differences in drug efficacy due to pharmacokinetic and pharmacodynamic characteristics; the complexity of drug dosing and administration in pediatrics; the fact that in most cases medications are used under indications different from those authorized and the intrinsic heterogeneity of the pediatric population. Recently, the Institute for Safe Med- ication Practices (ISMP) has edited a bulletin reflecting the need to adopt strategies for the prevention of ME in pediatrics [23].

      Despite its importance, few data are found in the literature on MR in the pediatric population. Coffey et al. described the implantation of an MR program in the pediatric population upon admission to a tertiary hospital center, and found the greatest number of treatment discrepancies to occur in polymedicated (4 or more drugs) patients and in those receiving antiepileptic medication [24]. In a study carried out in four hospitals in the United Kingdom to assess the efficacy of MR upon admission in 244 pediatric patients, 45% of the latter were seen to present at least one unintended discrepancy [25].

      A study carried out by the Department of Pediatrics of the Jordan University Hospital in 2018 involving 100 pediatric pa- tients conducted MR at the time of admission. At least one dis- crepancy was identified in 13% of the patients – the most com- mon situation being the omission of some drug [26]. In view of the repercussion of MR in relation to patient safety and its relevance in a high-risk group such as the pediatric population, the development of an MR model specifically targeted to pedi- atric patients is required, in the same way as has been done in adults, with the purpose of ensuring that all healthcare centers with a pediatric area incorporate such activities in order to im- prove the safety of these patients.

Aim of the review

      To adopt an MR model upon hospital admission in pediatrics in order to detect ME and thus improve the safety of pediatric patients.

To determine and analyze the discrepancies found between home treatment and the treatment prescribed upon admission; the drug groups implicated in such discrepancies; the back- ground diseases in those patients in which Reconciliation Errors (Res) are detected; and the pharmaceutical interventions car- ried out to resolve the discrepancies and thus avoid ME.

      To develop an algorithm for the identification of pediatric pa- tients with a priority indication of MR.

      All the procedures performed in this study, which involved human participants, were conducted in accordance with the ethical standards of the institutional and Spanish research com- mittees, with the 1964 Helsinki declaration and its later amend- ments, or with comparable ethical standards. Informed consent from patients was obtained at the clinical interview performed by the pharmacist in the Hospitalization Unit where the patient was admitted. The data were anonymized for analysis and only the researchers had access to the password-protected data- base. The participants were identified by a sequential numeri- cal code. The acquired information regarding participants was treated confidentially. The treatment of the data was carried out in accordance with Organic Law 03/2018 and European Regulation (EU) 2016/679, on the Protection of Personal data.

      The project was carried out in a pediatric hospital between January-November 2018 with the collaboration of the Prycon- sa^® Foundation. We included all patients admitted to the hos- pital during the study period, except those meeting any of the following exclusion criteria: psychiatric patients, subjects with an estimated duration of admission of under 24 hours, cases in which the clinical interview with the patient and/or caregiver was not possible, and patients in the emergency care depart- ment.

      The pharmacist in charge obtained a list of the patients ad- mitted in the last 24 hours based on the electronic prescrip- tion module of Farmatools^®. Data compilation was made on a continuous basis on working days from 9 a.m. to 2 p.m. The list was obtained on Mondays, taking into account the admissions of the last 72 hours.

      The Best Possible Medication History (BPMH) [15] was ob- tained for each patient included in the study. The BPMH in- cluded the complete medication of the patient throughout the healthcare process at both ambulatory (chronic medication) and hospital level (medication during admission). It also includ- ed the history referred to over-the-counter medications (public- ity products, herbal remedies, homeopathic or parapharmacy products, etc.), patient adherence to treatment, and drug al- lergies.

The information sources used to obtain the BPMH were

• Electronic case history (HCIS^®), with a detailed review of the available reports referred to patient hospital admission epi- sodes and surgeries or nursing care.

• Primary care clinical/prescription history (HORUS^®), review- ing the available reports referred to hospitalization episodes in other centers of the Community of Madrid, primary care re- ports, and active patient prescriptions

• Clinical interview with the patient or caregiver. The phar- macist visited the hospitalization unit to which the patient was admitted in order to conduct a clinical interview with the pa- tient or caregiver, asking questions referred to current home treatment, adherence to therapy and possible drug allergies or intolerances.

      The following variables were recorded for each reconciled patient: date of admission, date of reconciliation, age, gender, reason for admission, background disease, drug allergies or in- tolerances, number of home drugs (with polymedication being defined as the use of 4 or more drugs), drugs with a Narrow Therapeutic Index (NTI) (Table 2), phytotherapeutic or homeo- pathic treatments, patient pertaining to the Community of Madrid, updating of treatment prescribed in HORUS^®, discrep- ancies, number of discrepancies per patient, type of discrep- ancy, Pharmaceutical Intervention (PI), type of PI, resolution of discrepancy, drug implicated in the discrepancy, and drug class implicated in the discrepancy.

      The background diseases were classified as: pediatric com- mon diseases, neurological, oncohematological, mitochondrial and psychiatric disorders, asthma, autoimmune diseases, diges- tive, metabolic, endocrine, genetic and renal diseases, hemato- logical disorders and cystic fibrosis. The drug classes implicated in the discrepancy were registered according to the Anatomic, Therapeutic, Chemical (ATC) classification of the Spanish Medi- cines Agency (AEMPS).

      Once the BPMH of each reconciled patient was obtained, an analysis was made of the possible discrepancies between it and the treatment prescribed upon admission. The pharmacist compared the BPMH against the available treatment instruc- tions, taking into account the current clinical situation of the patient, the justification of the prescriber, and the drug treat- ment indications.

      The discrepancies were classified based on the medication reconciliation terminology and classification consensus docu- ment of the Spanish Society of Hospital Pharmacy as either jus- tified or unjustified the latter being taken to represent Recon- ciliation Error (RE) and therefore capable of leading to ME [27].

      The following were regarded as justified discrepancies: med- ical decision not to prescribe a drug or to modify its dose, fre- quency or administration route in accordance to the new clinical situation; medical decision to modify drug posology or adminis- tration route in accordance to the new clinical situation; intro- duction of a new medication justified by the clinical situation; therapeutic substitution according to the hospital center phar- macotherapeutic guide and therapeutic exchange programs. In turn, unjustified discrepancies (and therefore Res) were classi- fied into the following groups: omission, commission, different dose, administration route or frequency of a drug, duplicity, in- teraction, maintenance of a medication in a situation in which it is contraindicated, and incomplete prescription.

      Pharmaceutical Intervention (PI) represented the last step in the MR process, and the physician in charge was informed of the REs detected through the electronic prescription system. The most urgent or important PI was also reported verbally. Subsequently, confirmation of the resolution of the RE was made by reviewing the patient treatment order. Lastly, the data obtained in the MR process were entered in the MS Excel data- base designed for the study.

      The analysis of the data was performed using the Stata v.13 statistical package in the Methodology Unit of the Instituto de Investigación Princesa (IP). A descriptive analysis was made, with the calculation of central tendency (mean and/or median) and dispersion measures (Standard Deviation [SD] and/or per- centiles).

      During the 11 months of the study, MR was carried out in 1760 patients (45% females and 55% males), with a mean age of 7.9 ± 5, 2 years (range 12 days - 24 years) and a mean body weight of 28.9 ± 18.6 kg (range 3.2-114.2). Sixty percent of the patients had some background disease, the most common be- ing neurological disorders, present in 18% of the patients (Table 3).

      Most of the patient admissions were to medical units (60% versus surgical units in 40%).

      On analyzing the type of home treatment, 14% of the cases (237 patients) were considered to be polymedicated, with an average of 6 drugs per patient (range 4-15). Approximately the same number of patients (n=251) had some NTI drug included in their routine home treatment, and only 1.7% used phytother- apy or homeotherapy.

      Eighty percent of the study sample consisted of patients pertaining to the Community of Madrid, and of these, 39% did not have up to date information on their home medication in HORUS^®.

      In 592 patients we detected at least one discrepancy be- tween home treatment and the treatment prescribed upon admission (Table 4). The total number of discrepancies detect- ed was 830, of which 307 (37%) were justified and 523 (63%) unjustified (and thus represented RE), affecting 334 patients (56%). On analyzing the REs, most were seen to be due to omis- sion (77%), followed by different dose, administration route or frequency (17%) (Table 5). Thirty-two percent of the patients with RE suffered background neurological disease, fundamen- tally epilepsy, and 16% presented oncohematological disease (Table 6).

      Most of the patients with RE were admitted to medical units (56% versus surgical units in 43%).

      Of the patients in which REs were detected, 34% (n=112) were polymedicated, with an average of 6 drugs per patient (range: 4-13) and 30% (n=99) had some NTI drug in their home treatment. The main drug groups implicated in the detected REs corresponded to psycholeptics and psychoanaleptics (n=58), antiepileptic drugs (n=57) and systemic antibacterials (n=51) (Table 7).

      In order to avoid MEs, a total of 460 PIs were carried out. Of these, 76% aimed to start regular treatment that had been omitted upon admission. The degree of acceptance of the PIs on the part of the healthcare professionals was 72% (Table 8).

Table 1

Table 1: Drugs with a reconciliation time of under four hours [8].

Table 2

Table 2: Drugs with a narrow therapeutic index (Modified from AEMPS [28]).

Table 3

Table 3: Demographic and clinical characteristics of the study sample (n=1760).

Table 4

Table 4: Classification of the patients according to the presence or absence of discrepancies.

Table 5

Table 5: Classification of the discrepancies.

Table 6

Table 6: Demographic and clinical characteristics of the patient with reconciliation error (n=334).

Table 7

Table 7: Distribution of reconciliation errors according to drug class (ATC classification).

Table 8

Table 8: Pharmaceutical interventions.

      The results of our study show that 34% of the patients had at least one discrepancy in their treatment, and that 18% present- ed RE that could have resulted in ME. These figures are similar to those published by Abu Farha et al., with RE in 13% of their cases [26], and lower than those obtained by Huynh et al., who found 45% of their patients to have at least one unintended discrepancy [25]. The comparison of our results with those of other authors is complicated due to the fact that there are few studies in the literature on MR in the pediatric population, and the lack of uniformity of the terminology used. To the best of our knowledge, the present study represents the first published experience with the implantation of a medication reconciliation model in the pediatric population at national level. In order to avoid ME, a total of 460 PIs were carried out by a hospital phar- macist. This is consistent with the data published by other au- thors regarding increased patient safety thanks to the MR made by a hospital pharmacist [16].

      As in other studies of MR in both adults and the pediatric population [10,11,24,26], the most common RE was the omis- sion of some drug (39% of all REs), followed by incorrect pre- scription of drug dose, administration route or frequency (10% of all REs). One of the reasons for this could be the lack of up- to-date and easily accessible information regarding the home treatments of the patients at the time of admission to hospital. This is supported by our own experience, since 39% of the home treatment specifications of the reconciled patients pertaining to the Community of Madrid did not coincide with the information available in HORUS^®.

      In turn, we found a possible relationship between the num- ber of drugs used by the patient at home and the presence of RE upon admission – a larger number of drugs being associated to a greater probability of RE. These results are consistent with the published observations in adults, where polypharmacy was like- wise established as one of the main risk factors for RE [10,11].

      It is important to underscore that to the best of our knowl- edge, this is the first study to analyze RE according to the type of pediatric patient. In this regard, the main groups of patients in which REs were detected corresponded to those with neu- rological, oncohematological and severe respiratory diseases. Reconciliation errors were less frequently identified in other groups of patients such as those with autoimmune disorders, metabolic diseases, cystic fibrosis or non-oncological hemato- logical diseases. Due to the complexity of these patients and the importance of correct pharmacological treatment, we con- sider that patients of this kind should also be included among those amenable to reconciliation.

      In order to guarantee the safety of pediatric patients, we also analyzed the drugs most often implicated in RE. In this regard, the drug classes most commonly implicated in RE were antiepi- leptic drugs, psychoanaleptics and psycholeptics, and systemic antibacterials. These results are consistent with those of other studies in adult patients, where these same drug classes were seen to be those most often implicated in RE [10].

      The high percentage of PIs accepted by the medical team fol- lowing the detection of RE (76%) reflects the importance of MR for the prevention of ME, and thus for ensuring increased safety of pediatric patients during hospital admission.

      Our results show that REs are common in the pediatric popu- lation and that the implantation of an MR model upon admis- sion would allow us to detect them before they affect the pa- tient. Based on the results obtained, we attempted to identify the main characteristics of those pediatric patients at greatest risk of suffering RE, as well as the drug classes most often im- plicated in such errors. These data could help us to select the patients, assigning priority MR to those individuals that stand to benefit most or which could be most susceptible to ME, with a view to reinforcing their safety during hospital admission.

      On considering the background disease of the patient, MR should focus on chronic pediatric patients with neurological, oncohematological or severe respiratory disorders, and on those receiving regular treatment with antiepileptic drugs, psy- choleptics and/or psychoanaleptics and systemic antibacterials.

      It would be advisable to conduct a study to confirm that pa- tients with these characteristics effectively would be those de- riving most benefit from reconciliation processes.

      Among the limitations of our study, mention must be made of the lack of available resources for performing MR in all pa- tients admitted from 3 p.m. onwards; the lack of evaluation of the clinical significance of RE as done in other studies in the pediatric population [25]; and the impossibility of assessing the acceptance of some PIs because the patient was discharged be- fore verification of the resolution of the discrepancies.

      It also would have been interesting to have a reference or control group without PI, in order to compare the intervention versus standard treatment.

      A multicenter study with other pediatric hospitals would be advisable in order to more closely analyze patients with back- ground diseases that are infrequent in our center (e.g., heart disease), since such a study would generate a more represen- tative sample size allowing the drawing of more solid and ex- trapolatable conclusions.

      Medication reconciliation in pediatrics, in the same way as in the adult population, has been shown to be useful in the detec- tion of RE, contributing to prevent such errors from reaching the patient. The MR criteria in adults cannot be extrapolated to the pediatric setting; it therefore would be necessary to de- velop and validate an algorithm allowing the selection of those patients that stand to benefit most from MR, with the implanta- tion of this activity in all centers that treat the pediatric popula- tion.

      The authors would like to thank the Pryconsa Foundation, which financed the expenses necessary to carry out this project.