Validating the Chinese geriatric trigger tool and analyzing adverse drug event associated risk factors in elderly Chinese patients: A retrospective review.

OBJECTIVE: The aim was to evaluate the performance of the initial Chinese geriatric trigger tool to detect adverse drug events (ADEs) in Chinese older patients, to attempt to shorten this list for improving the efficiency of the trigger tool, and to study the incidence and characteristics of ADEs in this population.
METHODS: A sample of 25 cases was randomly selected per half a month from eligible patients who aged 60 years and older, hospitalized more than 24 hours, and discharged or died between January 1, 2015 and December 31, 2017 in West China hospital. A two-stage retrospective chart review of the included inpatients were conducted. ADEs were detected using a list of 42 triggers previously selected by an expert panel by means of a Delphi method. The number of triggers identified and ADEs detected were recorded and the positive predictive value (PPV) of each trigger was calculated to select the most efficient triggers. Several variables were recorded, including age, sex, number of diseases, length of hospital stay and so on, to analyze the risk factor of ADEs.
RESULTS: Among 1800 patients, 1646 positive triggers and 296 ADEs were detected in 234 (13.00%) patients. Older patients who were younger, had more medications, longer stays or more admission, and did not experience surgical operation more likely experienced ADEs. Triggers with PPV less than 5% were eliminated, which resulted in the upgraded version of Chinese geriatric trigger tool of 20 triggers with a PPV of 28.50%. This upgraded tool accounted for 99.66% of all ADEs detected.
CONCLUSIONS: The upgraded version of Chinese geriatric trigger tool was an efficient tool for identifying ADEs in Chinese older patients. Future, the trigger tool could be incorporated into routine screen systems to provide real-time identification of ADEs, thereby enabling timely clinical interventions.

Introduction

Patient safety data from all across the globe indicate that the burden of medication-related harm is very high. The Harvard Medical Practice Study showed that medication-related injuries were the most frequent cause of adverse events and were correlated with disabling injury in about 1% of all hospitalized patients [1]. Researchers have suggested that medication-related harms account for prolonged hospital stays, cause 100,000 deaths per year and cost as much as $10 to $150 billion in the United States annually [2,3]. For these reasons, the World Health Organization (WHO) has launched its third global patient safety challenge to promote and implement actions for improving medication safety and reducing the number of preventable adverse drug events (ADEs) [4].

Older patients are more likely to experience drug-related events and have higher ADE prevalence rates compared with other age groups due to multiple co-morbid illnesses, polypharmacy, difficulty monitoring prescribed medications, and age-related changes in pharmacokinetics and pharmacodynamics [5-7]. Older adults in the United States account for 35% of all hospital stays and 53.1% of hospital ADEs [8]. Therefore, the reduction of ADEs in these vulnerable patients has become a significant safety goal in various settings [9].

Currently, various methods for identifying ADEs have been proposed, including screening voluntary reports, mining administrative databases and reviewing patient claims and medical records [10,11]. However, the vast majority of ADEs go undetected by these traditional methods, and common detection techniques have not produced consistent approaches to measure harm [12,13]. To improve medication safety and achieve the objective of reducing the number of ADEs, health care organizations should have efficient and feasible surveillance strategies to measure ADEs and monitor the results of improvement interventions.

The Global Trigger Tool (GTT), which was developed by the Institute for Healthcare Improvement in 2003, is a low-resource option for detecting adverse events at hospitals [13]. By using “triggers” to guide medical record review, GTT is more efficient in identifying adverse events compared to traditional methods [11,14]. This tool can be used in clinical practice to track and assess adverse event rates. The GTT may also be integrated into health information technology to meet the demands of rapid and real-time identification of adverse events, enable timely interventions to mitigate adverse events, and determine the effectiveness of these interventions over time [13,15]. In addition, for different types of events, groups of people or clinical settings, such as drug, perinatal, pediatric, ambulatory care settings or mental health settings, specific sets of triggers can be customized [16-18].

In China, the National Center for Adverse Drug Reactions Monitoring has established the National Adverse Drug Reaction Monitoring System, a spontaneous reporting system, to report each adverse drug reaction or adverse drug event (ADR/ADE) and to improve the data quality management of ADR/ADE reports, a normative grading criterion based on the WHO criteria. Despite the availability of this surveillance strategy, the incidence and characteristics of ADEs in Chinese patients are largely unknown, especially for specific groups of people or medications. Therefore, based on existing triggers described in the literature and evaluated by an expert panel following a Delphi method, an initial trigger tool for older Chinese patients was developed [19]. The initial list included a total of 42 triggers divided into six categories (laboratory index, plasma concentration, antidote, clinical symptom, intervention and other) [19].

The aim of the present study was to validate the initial Chinese geriatric trigger tool in older patients in clinical practice; to improve the efficiency of the trigger tool by attempting to shorten this list in accordance with the result obtained; and to estimate the frequency and characteristics of ADEs in this population.

Methods

Study design and setting

This observational, retrospective study was conducted in the West China hospital of Sichuan University, a large, tertiary general teaching hospital in China. The West China hospital has 4300 beds and provides medical services to the west and south region of China [20]. There were 263, 700 inpatients, as well as 5.44 million outpatient and emergency patients in 2018; and about a quarter of all patients were geriatric patients [20]. This hospital uses electronic medical record (EMR) and bar code systems to document medication administrations. Ethics approval was obtained from the respective ethics committees at the West China Hospital of Sichuan University, China (2018–502). The institutional ethics committee waived the requirement of informed consent for this retrospective study, and all data used in this study is fully anonymized.

Eligible patients were aged ≥60 years by the official Chinese definition [21] who had a hospital stay of no less than 24hours and had been discharged or died between January 1, 2015 and December 31, 2017. Patients who were admitted to the psychiatric, rehabilitation, ambulatory surgery, and integrated traditional/ western medicine ward were excluded. A literature review revealed that the rate of geriatric inpatient ADEs is about 20–25% [22,23]. According to calculation, the sample size (N) was set as 600 cases per year [24]. After sorting by date of discharge or died, a sample of 25 cases was randomly selected from eligible patients per half a month, for a total of 1800 cases [13]. Charts for review were randomly selected from the list of eligible patients using the randomization function found at https://www.random.org/sequences/.

Triggers

The development of the set of triggers that was used in the study has been previously reported [19]. Briefly, 51 triggers that had been identified through a detailed literature review were evaluated by an expert panel for appropriateness for geriatric patients by two-round Delphi method. The developed tool included a total of 42 triggers that were organized in the following six categories: 15 laboratory indices, 6 plasma concentrations, 13 antidotes, 6 clinical symptoms, one intervention and one other (S1 Table).

Records review

The researchers who completed this study were trained on its methodology through participation in a similar, previous study [21]. In the current study, a two-stage review process for medical records was conducted. In the first stage, two trained clinical pharmacists (Hu and Qin) independently reviewed each medical record for the presence of any of the triggers with a limit of no more than 20 min per chart [13]. The following sections of the charts were reviewed: medical progress notes, nursing flow sheets, medication orders, and laboratory data. Each identified trigger was recorded for further chart analysis to determine whether an associated ADE had occurred. In the second stage, one physician reviewed all the medical records with identified triggers from the first stage to determine the presence of an ADE and assign its respective category and severity. If there was a disagreement, the final decision was made based on a consensus at the study group meetings.

Besides the identified triggers and detected ADEs, the following variables were also documented: age, sex, number of diseases, length of hospital stay, number of admissions in the previous 1-year, number of medications, number of doses, and surgeries. The severity of each ADE was evaluated using the National Coordinating Council for Medication Error Reporting and Prevention Index (NCC MERP) [25]. We focused on ADEs that cause actual patient harm. Therefore, only ADEs from categories E to I were recorded: E (temporary harm to the patient requiring intervention), F (temporary harm to the patient and requiring initial or prolonged hospitalization), category G (permanent patient harm), H (intervention required to sustain life), and I (patient death). The researchers determined whether the ADEs could have been prevented using the Six-Item Screener, in which 1 indicates “definitely not preventable” and 6 indicates “definitely preventable”. The ADEs with ratings greater than 4 (i.e., more than 50% likelihood preventable) were classified as preventable [26].

The positive predictive value (PPV) of each trigger was calculated using the number of identified ADEs related to this trigger divided the number of times the trigger. The PPV for the overall trigger tool was also calculated. Finally, based on a similar study and our previous study, the triggers that were found to have a rate higher than a preestablished PPV cutoff (5%) were retained in the final tool [24,27].

Statistical analysis

The SPSS 25.0 software was used to analyze data. Descriptive statistics were calculated for patients and ADE characteristics. The following rates were calculated: ADEs per 100 admissions, ADEs per 1000 patient days, ADEs per 1000 doses, and ADEs per 1000 medications [13]. Categorical variables were summarized using frequency counts and percent, and continuous variables were presented as means with standard deviations (SD) and medians with ranges. Comparisons between groups were made using the nonparametric Mann-Whitney U test for continuous variables, and the χ2 test for categorical variables. Stepwise logistic regression was performed for variables which associated with diagnoses at a significance level of P<0.1 in univariate analysis [28]. Any variable significant at a level of 0.05 after regression was reported as an independent risk factor for ADEs [28]. Multicollinearity diagnosis was performed by variance inflation factors (VIF). The variables with a VIF of more than 4 were removed [29].

Results

Patients characteristics

A total of 1800 randomly selected patients were reviewed. Among these patients, 746 (41.44%) were female and the mean age was 69.84 years (range 60 to 101). The mean length of hospital stays varied broadly, so that the median number of lengths of stay was 8 days (range 1 to 89). The number of medications taken per patient also varied broadly with a median of 6 (range 1 to 37), and the median number of doses per patient was 39.5 (range 1 to 1731). Of the 234 (13.00%) patients with ADEs, 47 patients suffered from more than one ADE (Table 1). According to studies in Chinese hospitals, four of the most common diseases in older Chinese people are neoplasms and diseases of the circulatory, digestive, and respiratory systems [30-32], which are similar to our study’s findings. The proportion of each diseases found in our sample was lists in S2 Table.

Table 1

Patients characteristics.

Characteristics	Total (n = 1800)	Patients with ADEs (n = 234)	Patients without ADEs (n = 1566)	OR	95%CI	P
Sex
Male	1054	154	900	1.424	1.068–1.900	0.019
Female	746	80	666
Age (years)
60–74	1326	189	1137	1.585	1.124–2.234	0.009
74-	474	45	429
Intensive care units
Yes	54	13	41	2.188	1.154–4.248	0.017
No	1746	221	1525
Admission in the previous 1-year
Yes	814	165	649	3.379	2.507–4.554	<0.001
No	986	69	917
Allergic history
Yes	193	24	169	0.945	0.601–1.484	0.91
No	1607	210	1397
Surgery performed
Yes	660	32	628	0.237	0.161–0.348	<0.001
No	1140	202	938
Type of admission
Elective	1507	187	1320	0.741	0.524–1.050	0.11
Emergent	293	47	246
Method of admission
Wheel chair, cart or other assistance	340	56	284	1.420	1.024–1.969	0.039
On foot	1460	178	1282
Treatment outcome
Improve or cured	1584	197	1387	0.687	0.468–1.009	0.066
Did not improve or died	216	37	179
Antibacterial use
Yes	577	80	497	1.117	0.836–1.494	0.454
No	1223	154	1069
Chinese patent medicine use
Yes	223	37	186	1.393	0.950–2.044	0.09
No	1577	197	1380

The univariate analyses showed there were no significant differences in allergic history, type of admission, treatment outcome, antibacterial use and Chinese patent medicine use (P > 0.05), whereas significant differences were identified in sex, age, length of stay, number of medical diagnoses, number of admission in the previous 1-year, number of medication and doses, method of admission, intensive care units and surgeries between patients with and without ADEs (P ≤ 0.05) in Tables 1 and 2.

Table 2

Mann-Whitney U test result of risk factors.

Characteristics	Total (n = 1800)	Patients with ADEs (n = 234)	Patients without ADEs (n = 1566)	Z	P
Age (years)	69.84±8.14	67.95±7.55	70.12±8.16	-4.125	<0.001
Length of stay	10.19±8.94	15.19±12.28	9.44±8.07	-7.679	<0.001
Number of medical diagnoses	4.70±3.31	5.73±3.74	4.55±3.21	-5.237	<0.001
Number of admissions in the previous 1-year	1.34±3.17	2.33±3.07	1.20±3.16	-10.001	<0.001
Drugs per patient	7.19±5.71	10.00±6.26	6.77±5.51	-7.938	<0.001
Doses per patient	92.63±140.74	157.91±228.43	82.87±119.45	-5.521	<0.001

ADEs and risk factors

Among the 234 patients with ADEs, 296 ADEs were identified, including 25 preventable ADEs. Two hundred eighty-two ADEs (95.27%) occurred during hospital stays, and 14 (4.73%) pre-existed as the reasons for the hospital admission. Two hundred thirty-two ADEs (78.38%) were determined to be harm category E of NCC MERP; 50 (16.89%) were category F; 13 (4.39%) were category H; and 1 (0.34%) was category I (Table 3). The calculated rate of ADEs was 16.44 per 100 admissions and 16.14 per 1000 patient days, 22.60 per 1000 medications, and 1.77 per 1000 doses.

Table 3

Types of adverse drug events.

Organ /System	ADE	Total ADEs		Preventable ADEs
		n	%	n	%
Metabolism and nutrition	Hypokalemia	4	1.35%	2	8.00%
	Hyperkalemia	2	0.68%	2	8.00%
	Hypoglycemia	4	1.35%	2	8.00%
Liver and biliary system	Hepatotoxicity/ Transaminase disorder	25	8.45%	0	0.00%
Urinary system	Nephrotoxicity/ Creatinine disorder	3	1.01%	1	4.00%
	Urinary Retention	1	0.34%	0	0.00%
Infectious	Candidiasis	2	0.68%	1	4.00%
	Infection of the upper respiratory tract	1	0.34%	0	0.00%
Musculoskeletal	Myalgia	1	0.34%	0	0.00%
Immune system	Allergy	19	6.42%	1	4.00%
Constitutional symptoms	Fever	10	3.38%	0	0.00%
	Weakness	5	1.69%	0	0.00%
	Pain	1	0.34%	0	0.00%
	Cold sweating	1	0.34%	0	0.00%
Central and peripheral nervous system	Dizziness	4	1.35	0	0.00%
	Sleepiness	2	0.68%	0	0.00%
	Tremor	1	0.34%	0	0.00%
Gastrointestinal	Constipation	9	3.04%	0	0.00%
	Diarrhea	22	7.43%	9	36%
	Nausea	64	21.62%	0	0.00%
	Anorexia	10	3.38%	0	0.00%
	Vomiting	22	7.43%	0	0.00%
	Acute gastric mucosal Injury	1	0.34%	1	4.00%
	Abdominal distension	1	0.34%	0	0.00%
Cardiovascular	Hypotension	6	2.03%	1	4.00%
	Palpitation	2	0.68%	0	0.00%
	Bradycardia	1	0.34%	0	0.00%
	Vascular headache	1	0.34%	0	0.00%
Hematologic	Hemorrhage	21	7.09%	5	20%
	Leukopenia	36	12.16%	0	0.00%
	Thrombocytopenia	13	4.39%	0	0.00%
	Hemoglobin decline	1	0.34%	0	0.00%
Total		296	100.00%	25	100%

Multicollinearity diagnostic results showed that the VIF of doses per patient was larger than 4, which indicated doses per patient had collinearity with other factors (S3 Table). Logistic regression results showed that the significant factors associated with the occurrence of ADEs were age, length of stay, surgery, number of medication and admissions (P < 0.05) in Table 4 (The complete result was showed in S4 Table).

Table 4

Stepwise logistic regression results of the occurrence of ADEs.

Variables	B	SE	Wald	P	Exp(B)	95%CI
Sex (Female)	0.293	0.156	3.518	0.061	1.340	0.987–1.820
Age	-0.046	0.010	21.184	0.000	0.955	0.937–0.974
Length of stay	0.043	0.009	24.595	0.000	1.044	1.027–1.062
Intensive care units	0.647	0.378	2.934	0.087	1.910	0.911–4.006
Number of admissions in the previous 1-year	0.067	0.018	13.815	0.000	1.069	1.032–1.107
Surgery performed	-1.252	0.205	37.301	0.000	0.286	0.191–0.427
Method of admission (On foot)	-0.397	0.206	3.703	0.054	0.673	0.449–1.007
Number of medications per patient	0.047	0.015	9.787	0.002	1.048	1.018–1.080
Constant	0.443	0.683	0.420	0.517	1.557

Among the 42 triggers, 34 (80.95%) were positive and 23 (54.76%) were associated with ADEs. The result of triggers was divided into two blocks (Tables 5 and 6) to summarize the outcomes of the triggers that had PPVs more than 5% and of all other triggers with PPVs less or equal than 5% (cutoff point). A total of 1646 triggers were detected, and 343 were related to ADEs (an ADE could be identified by one or more triggers). The overall PPV of the Chinese geriatric trigger tool was 20.84%. A wide variability was found in the ADEs detected and the PPVs within the six categories. The triggers of laboratory index and antidotes allowed for more ADEs to be identified, but the plasma concentration triggers identified fewer ADEs.

Table 5

Prevalence of selected triggers and ADEs.

No.	Selected Triggers, n		Total ADEs		Preventable ADEs
			n	PPV, %	n	PPV, %
Laboratory index
1	PTT > 100s	2	2	100.00%	1	50%
2	INR > 5	2	2	100.00%	1	50%
3	Glucose < 2.8mmol/L	6	4	66.67%	2	33.33%
4	Rising BUN or serum creatinine greater than 2 times baseline	10	2	20.00%	1	10%
5	ALT (or AST) ≥3 ULN and / or ALP≥2 ULN and T-BIL > 2UNL (can have abnormal INR)	51	31	60.78%	0	0.00%
6	PLT < 75×109/L	39	13	33.33%	0	0.00%
7	WBC < 3.0×109/L	52	37	71.15%	0	0.00%
8	Decrease of greater than 25% in hemoglobin or hematocrit	9	4	44.44%	2	22.22%
9	K+ < 3.5mmol/L	109	6	5.50%	2	1.83%
10	K+ > 5.5mmol/L	13	3	23.08%	3	100.00%
Antidote
11	Antiallergic	70	16	22.86%	1	1.43%
12	Anti-emetic	536	101	18.84%	0
13	Antidiarrheal	31	16	51.61%	7	22.58%
14	Laxative	137	14	10.22%	1	0.73%
15	Transfusion or use of blood products	28	5	17.86%	2	7.14%
Clinical symptom
16	Over sedation/hypotension	18	10	55.56%	1	5.56%
17	Rash	19	14	73.68%	0	0.00%
18	Heart rates <60/min	3	2	66.67%	0	0.00%
Intervention
19	Abrupt medication stops	40	40	100.00%	4	10.00%
Other
20	Others ADEs (ADEs not related to one of the triggers listed above)	18	18	100.00%	2	11.11%
Subtotal		1193	340	28.50%	30	2.51%

PTT, partial thromboplastin time; INR, international normalized ratio; BUN, blood urea nitrogen, ALT, alanine aminotransferase; AST, aspartate aminotransferase; T-BIL, total bilirubin; PLT, platelets; WBC, white blood cells; K+, potassium; UNL, upper limit of normal; PPV, positive predictive value.

Table 6

Prevalence of excluded triggers and ADEs.

No.	Not-selected Triggers, n		Total ADEs		Preventable ADEs
			n	PPV, %	n	PPV, %
Laboratory index
1	HGB > 170g/L(man), > 150g/L(woman)	0	0	—	0	—
2	Ca2+ > 2.62 mmol/L	4	0	0.00%	0	0.00%
3	TSH < 0.27 mU/L or FT4 > 22.40 pmol/L	1	0	0.00%	0	0.00%
4	TSH > 4.2mU/L or FT4 < 12.0 pmol/L	13	0	0.00%	0	0.00%
5	Clostridium difficile positive	0	0	—	0	—
Plasma concentration
6	Digoxin > 2 ng/ mL	0	0	—	0	—
7	Gentamicin or Tobramycin levels peak > 10mg/L, trough > 2mg/L	0	0	—	0	—
8	Cyclosporin > 300μg/mL	0	0	—	0	—
9	Theophylline > 20mg/L	0	0	—	0	—
10	Tacrolimus > 20 ng/mL	0	0	—	0	—
11	Voriconazole levels > 5.5mg/L	0	0	—	0	—
Antidotes
12	Vitamin K	29	1	3.45%	1	3.45%
13	Romazicon (Flumazenil)	117	0	0.00%	0	0.00%
14	Naloxone (Narcan)	4	0	0.00%	0	0.00%
15	50% glucose	0	0	—	0	—
16	Protamine	6	0	0.00%	0	0.00%
17	Epinephrine	230	2	0.87%	0	0.00%
18	Glucose injection and regular insulin	30	0	0.00%	0	0.00%
19	Insulin (regular insulin or insulin analogue) used in non-diabetics	2	0	0.00%	0	0.00%
Clinical symptoms
20	Dehydration	1	0	0.00%	0	0.00%
21	Psychosis	1	0	0.00%	0	0.00%
22	Respiratory rates < 12 /min	15	0	0.00%	0	0.00%
Subtotal		426	3	0.70%	1	0.23%

HGB, hemoglobin; Ca2+, calcium; TSH, thyroid stimulating hormone; FT4, free thyroxine.

The 20 triggers with PPVs accounting for more than 5% of the total were selected to become the upgraded version of Chinese geriatric trigger tool, which increased overall PPV increase to 28.50%. There was only one ADE not identified by the 20 triggers. The upgraded version of the trigger tool accounted for 99.66% of all the ADEs and 100% of the preventable ADEs (Table 5).

Discussion

Clinicians should prioritize actions that reduce incidence of avoidable harm caused by medication in their older patients. Through integrating a literature review of existing triggers with a Delphi process, we have developed an initial list of 42 triggers for detecting ADEs in Chinese geriatric inpatients [19]. Conducting a pilot-testing of this 42-trigger tool in 1800 Chinese older patients led to the identification of 13.00% of older patients with at least one ADE, and the initial list was shortened based on the results obtained. Through use of the cutoff PPV value, less robust triggers were removed and the efficiency of the trigger tool was improved. Sixteen triggers with a PPV above 20% allowed for the detection of only 62.39% of all ADEs. The triggers with a PPV above 10% allowed for the detection of 97.38% of all ADEs but the hypokalemia trigger, which could not be substitute, would not be included. Therefore, the 20 triggers with a PPV above 5% were included into the upgraded version of the Chinese geriatric trigger tool [24,27], which covers the vast majority of ADEs, makes detection of ADEs easier and allows clinicians to identify ADEs in real time. The number of triggers in the upgraded version of Chinese geriatric trigger tool was smaller compared with other trigger tools for measuring ADEs in specific populations [6,22,27], and larger when compared to the general trigger tool for ADEs [13].

Twenty-two triggers from the initial list were eliminated based on the PPV cutoff. It is worth noting that all plasma concentration triggers were eliminated. This trigger category showed a low response rate because there was no patient conducted therapeutic drug monitoring (TDM). We attribute this to the fact that TDM has not yet become widespread in China, and plasma concentration monitoring has not served as a routine monitoring index in most Chinese older inpatients [19]. Therefore, we concluded that these triggers might had an acceptable applicability for the small numbers of patients undergoing TDM and that hospitals could add these triggers to the EMR in accordance with particular objectives or medication in the future [33]. The triggers that had a strong correlation with anesthesia did not allow for identification of any ADE or only detected one ADE. For example, flumazenil or naloxone was used to discontinue the induction and maintenance of anesthesia, but their use did not detect any ADEs. In addition, as China undergoes marketization and privatization, which poses numerous doctor-patient social problems such as a trust crisis [34], some emergency antidotes such as vitamin K or protamine are being used early to avoid medical disputes, although their use did not represent the occurrence of ADEs. Finally, other triggers, such as respiratory rates < 12 /min or Ca2+ > 2.62 mmol/L also did not identify any ADEs. These findings indicate the necessity of evaluating the set of triggers for use in real clinical practice.

In our study, the incidence of geriatric ADEs was found to be similar to or lower than those found in other studies [22,23,27], which might reflect variations in local practices and study participants. Among all ADEs identified in this study, most were identified as temporary harm to the patient, findings that were similar to other studies [35,36]. We found that the older patients who experienced ADEs received more medications during their hospitalization and had longer stays, as found in previous studies [22,27,35]. This shows that the number of medications taken is an important risk factor for ADEs and underscores the need to prioritize actions to benefit this especially vulnerable population.

In addition, ADEs might more likely be found in the older patients who were younger, did not experience surgeries and had had admission within the previous 1-year. To our knowledge, there are no studies indicating that the older elderly was more likely to experience ADEs [22,27,35,37]. The younger elderly was likely to experience ADEs because they more likely to receive high-risk medications. The older patients with a greater number of admissions in our study were mostly diagnosed with malignant tumors, kidney disease, diabetes and other chronic diseases. These disorders are often treated using specialized medications including chemotherapeutic drug, anticoagulants, non-steroidal anti-inflammatory drugs, and systemic corticosteroids, all of which have been shown to be high risk for ADEs [37]. On the contrary, because surgical patients were often only administered intravenous fluid therapy during surgery, these patients were administered fewer high-risk drugs, and therefore had a lower risk of ADEs.

There are several limitations in the present study that are inherent to trigger tool methodology. First, ADE detection was based solely on a retrospective review of the medical charts. The results were dependent on the quality of the documentation, which varied across different departments and doctors. Second, though the last trigger was “other ADEs”, triggers were limited in number and scope, and therefore they could not capture all ADEs. Third, we were limited in using only one hospital and its respective scope of treatment. There is still room for improvement in the upgraded version of the Chinese geriatric trigger tool, as other Chinese hospitals could customize this trigger tool according to their unique objectives and select the triggers that may be most useful at any given time for surveillance and for guiding system-level interventions such as those focused on identifying ADEs associated with a particular drug or drug group.

Conclusion

This study which included a large sample to validate the Chinese geriatric trigger tool and investigate ADEs in Chinese older patients. Despite the limitations of this study, the upgraded version of the Chinese geriatric trigger tool has been validated to an efficient list for identifying ADEs in older patients. In our study, more than 10% of the older inpatients experienced at least one ADE, and most of these experiences caused temporary harm. The most significant factors associated with ADEs included age, the number of medications administered, the length of stay, the number of admissions and whether the patient underwent surgery. In the future, the trigger tool could be incorporated into routine screening systems to provide real-time identification of ADEs, thereby enabling initiation of timely clinical interventions.

Initial Chinese geriatric trigger tool.

(DOCX)

Click here for additional data file.

Proportion of diseases among elderly patients.

(DOCX)

Click here for additional data file.

Multicollinearity diagnostic result.

(DOCX)

Click here for additional data file.

The complete result of stepwise logistic regression.

(DOCX)

Click here for additional data file.

(RAR)

Click here for additional data file.

31 Jan 2020

PONE-D-19-21999

Validating the Chinese geriatric trigger tool and analyzing adverse drug event associated risk factors in elderly Chinese patients: A retrospective review

PLOS ONE

Dear Mr Xu,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

All 3 reviewers independently concluded there were major revisions required. The reviewers have provided detailed instructions which we invite you to address. There are concerns around structure, language and, most importantly, statistics. We would appreciate receiving your revised manuscript by Mar 16 2020 11:59PM. When you are ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter.

To enhance the reproducibility of your results, we recommend that if applicable you deposit your laboratory protocols in protocols.io, where a protocol can be assigned its own identifier (DOI) such that it can be cited independently in the future. For instructions see: http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). This letter should be uploaded as separate file and labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. This file should be uploaded as separate file and labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. This file should be uploaded as separate file and labeled 'Manuscript'.

Please note while forming your response, if your article is accepted, you may have the opportunity to make the peer review history publicly available. The record will include editor decision letters (with reviews) and your responses to reviewer comments. If eligible, we will contact you to opt in or out.

We look forward to receiving your revised manuscript.

Kind regards,

Katie MacLure, PhD, MSc (dist)., BSc (1st)

Academic Editor

PLOS ONE

Journal Requirements:

When submitting your revision, we need you to address these additional requirements:

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at http://www.plosone.org/attachments/PLOSOne_formatting_sample_main_body.pdf and http://www.plosone.org/attachments/PLOSOne_formatting_sample_title_authors_affiliations.pdf

2. In the ethics statement in the manuscript and in the online submission form, please provide additional information about the patient records used in your retrospective study. Specifically, please ensure that you have discussed whether all data were fully anonymized before you accessed them and/or whether the IRB or ethics committee waived the requirement for informed consent. If patients provided informed written consent to have data from their medical records used in research, please include this information.

3. Thank you for stating the following in the Acknowledgments Section of your manuscript:

"The authors would like to thank all the researchers for their participation in the study.

The authors also would like to thank the Cadre Health Care Committee of Sichuan (2018-103), which approved the planning, execution and analysis as proposed in the grant application.

Funding: The Cadre Health Care Committee of Sichuan supported this study."

We note that you have provided funding information that is not currently declared in your Funding Statement. However, funding information should not appear in the Acknowledgments section or other areas of your manuscript. We will only publish funding information present in the Funding Statement section of the online submission form.

Please remove any funding-related text from the manuscript and let us know how you would like to update your Funding Statement. Currently, your Funding Statement reads as follows:

"The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript."

4. Please amend your manuscript to include your abstract after the title page.

5. Please include captions for your Supporting Information files at the end of your manuscript, and update any in-text citations to match accordingly. Please see our Supporting Information guidelines for more information: http://journals.plos.org/plosone/s/supporting-information.

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Partly

Reviewer #2: Partly

Reviewer #3: Partly

**********

2. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: I Don't Know

Reviewer #2: No

Reviewer #3: N/A

**********

3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: Yes

Reviewer #3: No

**********

4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: No

Reviewer #2: Yes

Reviewer #3: No

**********

5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: Abstract:

The abstract is well-organised with the necessary information. Comments:

-Avoid using terms such as elderly by adopting term such as older adults

Methods: Describe how 1800 patients were selected

Conclusion: Why a shorter upgraded version ?

-Language use such as The aim of the study..

To attempt to shorten the list

Was calculated to use to select ..

Led to detect

METHODS

Sample size: Cite the calculation method: The sample size 67 was expanded to 600 cases per year.

Please explain what is subheading “Study Source” means

Please explain how this is possible: A sample of 25 cases was randomly selected from 74 eligible patients per half a month, for a total of 1800 cases

Unclear statement: There was 84 one “other ADEs” category, which comprised ADEs not related to one of the triggers listed above).

I understood the tool has been described in this article by the same authors: Hu Q, Qin Z, Zhan M, Wu B, Chen Z, Xu T. Development of a trigger tool for the detection of adverse drug events in Chinese geriatric inpatients using the Delphi method. Int J Clin Pharm 2019; 28. doi: 10.1007/s11096-019- 304 00871-x

-Is the sample used in this study considered as part of the presented data in the published article?

Conclusion: Try to avoid over-generalisation:

To our knowledge, this was the first study which included such a large sample to validate the Chinese 236 geriatric trigger tool and investigate ADEs in elderly Chinese elderly patients.

Conflict of interest in term of promoting the hospital unnecessarily: This hospital has been ranked second for 10 consecutive years in the composite index ranking of Chinese hospitals.

Reviewer #2: The authors have validated previously-developed prediction model for adverse drug events (ADEs). The model was upgraded by reducing the predictors (the triggers of ADEs). Higher positive predictive value was the goal of this study. I have several concerns before it can be considered for publication.

Major issues:

1. Line 55~67: Please be straightforward in ‘Study Design and Setting’. The commonly-expected information are the study design, name of hospital, country, and period of data extraction. The period information is still unavailable. No need to explain sample size estimation in this study that was already conducted, except for those subset of comparison that has no significant difference.

2. Line 123~124: Why did the authors choose significance level P<0.1, instead of P<0.05?

3. Line 81~83, 99~100; Table 1, 3~5: Please make a single table showing the trigger, unadjusted odds ratio (OR) with 95% confidence interval (CI) and P-value, adjusted odds ratio (aOR) with the same attributes, and positive predictive value (PPV) of each trigger. The OR is taken from a univariate logistic regression for each trigger while the aOR is taken from multivariate logistic regression including each trigger with variables of patient characteristics. There were a lot of significant differences in patient characteristics between patients with and without ADEs, while those were not variables of interest. These may affect association between each trigger and the outcome. Since this study aim to reduce the triggers from the original tools, only triggers with significant aOR should be included for the final model. Then both original and upgraded trigger tools should be fitted using the same configuration of multivariate logistic regression. The patient characteristics should be forced into the models either those using original or upgraded triggers. No predictor selection should be applied for the model using original triggers. This feature/predictor selection method may or may not be applied to the model using upgraded triggers. However, it is suggested to use the selection method if the authors want to reduce the triggers even more. It is also suggested to compare selection method (forward vs. backward vs. stepwise) since stepwise selection do not always give the best prediction model compared to other selection method. In the end, the PPV using all triggers can be calculated from the predicted outcomes of the models using either original or upgraded triggers. Meanwhile, the PPV should be calculated from univariate logistic regression including one trigger and patient characteristics without any selection method.

Minor issues:

Typos, grammatical errors, potentially-misleading statistical report, inadequate arguments in discussion, etc. are found in the manuscript. Please correct those. For example:

All lines: Please put superscripted citation after the punctuation (if available).

In line 69: Do you mean ‘Data Source’?

In line 70~71: … who had had a hospital stay of more than 24hours … (double ‘had’; miss the space in ’24hours’)

In line 74: The number of cases is a part of results. Please do not mention it in Methods.

In line 84: … of the triggers listed above) --> there is no pair of the parentheses?

Line 117: Do we really need to mention the widely-used Microsoft Excel 2016? What was the critical contribution of this software for hypothesis testing?

In line 137: … and respiratory systems 29-31, results which are similar … --> … and respiratory systems,29-31 which are similar …

In line 241: … underwent surgery.. (double punctuation marks)

Reviewer #3: In this manuscript, the authors aimed to evaluate the performance of the Chinese geriatric trigger tool to detect adverse drug events (ADEs) and identify the ADE-risk factors in elderly Chinese patients. The study is interesting, however, a number of issues should be addressed before its publication.

Major issues:

1. In table 1, the authors did chi-square test to evaluate the inter-group difference for categorical variables (Table 1). Can the author explain why the age grouping at 74 and 90? In addition, since the number of people over 90 is small, it might be better to merge 74-90 and 90- into one group. Otherwise, the calculated test value might be biased.

2. The authors used the stepwise logistic regression to find the risk factors for ADEs and used variance inflation factors (VIFs) to identify collinearity and remove the variables with VIF > 4. According to the author's description "stepwise logistic regression was performed for variables associated with diagnoses at a significance level of p < 0.1", there should be 13 variables (7 categorical variables and 6 continuous variables) in the initial model. It would be better to represent the process of stepwise logistic regression in the Appendix and tell the readers which variables were removed due to collinearity.

3. The results described in lines 139-143 were inconsistent with Table 3.

4. The explanation of triggers in lines 79-84, lines 159-164, and tables 4 & 5 were confusing. According to tables 4 & 5, there were 14 laboratory indices, and 7 plasma concentrations, which were different from lines 82-83. It was not clear how the authors separate 42 triggers into Table 4 and Table 5, as well as what the definition of positive was.

Minor issues:

1. There were some errors in table 1, for example,

# of patients with ADEs and # of patients without ADEs;

total number for ICU (Yes vs No) and patients without ADEs for ICU (Yes vs No.)

**********

6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: No

Reviewer #3: No

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files to be viewed.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email us at figures@plos.org. Please note that Supporting Information files do not need this step.

9 Mar 2020

1) Reviewer #1:

-Abstract: Conclusion: Why a shorter upgraded version?

Answer: According to the previous practice and expert opinion, some new triggers had been added to the initial list. We thought these triggers could cover most of ADEs. We thought about adding new trigger that were common in the study. The efficiency of the new triggers was inaccessible, because the denominators are unknow and the PPV of new triggers could not been obtain. In addition, too much triggers are not a good thing, the PPV of the tool may low to led to decrease efficiency. Therefore, the aim was shortening the trigger tool in this study.

-Is the sample used in this study considered as part of the presented data in the published article?

Answer: No. The ample used in this study is bigger than published article (25 per half month in this study; 20 per half month in published article), so we conducted a re-sampling with computer.

2) Reviewer #2:

-Line 123~124: Why did the authors choose significance level P<0.1, instead of P<0.05?

Answer: Because the interplay of variables, the results of univariate analysis were different to the multivariate analysis. Therefore, in general choosing significance level P<0.1 or P<0.15 was to avoid missing some important risk factors.

-3. Line 81~83, 99~100; Table 1, 3~5: Please make a single table showing the trigger, unadjusted odds ratio (OR) with 95% confidence interval (CI) and P-value, adjusted odds ratio (aOR) with the same attributes, and positive predictive value (PPV) of each trigger. The OR is taken from a univariate logistic regression for each trigger while the aOR is taken from multivariate logistic regression including each trigger with variables of patient characteristics. There were a lot of significant differences in patient characteristics between patients with and without ADEs, while those were not variables of interest. These may affect association between each trigger and the outcome. Since this study aim to reduce the triggers from the original tools, only triggers with significant aOR should be included for the final model. Then both original and upgraded trigger tools should be fitted using the same configuration of multivariate logistic regression. The patient characteristics should be forced into the models either those using original or upgraded triggers. No predictor selection should be applied for the model using original triggers. This feature/predictor selection method may or may not be applied to the model using upgraded triggers. However, it is suggested to use the selection method if the authors want to reduce the triggers even more. It is also suggested to compare selection method (forward vs. backward vs. stepwise) since stepwise selection do not always give the best prediction model compared to other selection method. In the end, the PPV using all triggers can be calculated from the predicted outcomes of the models using either original or upgraded triggers. Meanwhile, the PPV should be calculated from univariate logistic regression including one trigger and patient characteristics without any selection method.

Answer: The unadjusted odds ratio (OR) with 95% confidence interval (CI) and P-value have been showed in Table 1; positive predictive value (PPV) of each trigger have been showed in Tables 5 and 6.

These triggers are derived from clinical logic to flag medical records, which alerts reviewers to initiate further in-depth investigations regarding the patient’s record to determine the presence or absence of an adverse event; for example, a trigger is a value of blood glucose lower than 28 mg/dL in a patient with oral antidiabetic or insulin, which may alert professionals to perform a more detailed record review for evidence that the patient has an associated hypoglycemia. This method can be used in practice to track and assess ADE rates. The triggers and risk factor were two different things. The characteristics of patients was the risk factor of ADEs, which could be used to the prediction of ADE. However, the triggers were just to use to detect, but not used to the prediction of ADE. The “Preventable ADEs” meant the ADEs that were preventable which judged by the research team, but not meant the trigger could predict ADE.

PPVs = number of identified ADEs/ number of times the trigger. Therefore, the PPV was just the indicators of evaluating suitability for trigger, but had nothing to do with patients with or without ADEs. So. the OR and aOR of PPV could also not be calculated, and the aOR could not the inclusion criteria of triggers. We also could not conduct logistic regression to PPV% of each trigger. Because the results of OR and logistic regression of triggers could not get, the selection trigger was based on a similar study and our previous study, the triggers that were found to have a rate higher than a preestablished PPV cutoff (5%) were retained in the final tool.

3) Reviewer #3:

-According to the author's description "stepwise logistic regression was performed for variables associated with diagnoses at a significance level of p < 0.1", there should be 13 variables (7 categorical variables and 6 continuous variables) in the initial model. It would be better to represent the process of stepwise logistic regression in the Appendix and tell the readers which variables were removed due to collinearity.

Answer: the “Admission in the previous 1-year” and “Number of admissions in the previous 1-year” were the same thing. So, there were 12 variables (6 categorical variables and 6 continuous variables) in the initial model. The complete results of stepwise logistic regression and collinearity were showed in S3 and S4 Tables

-It was not clear how the authors separate 42 triggers into Table 4 and Table 5, as well as what the definition of positive was.

Answer: The triggers which PPVs%＞5 were showed in Table 4 (now is Table 5), and these triggers were retained in the final tool. The triggers which PPVs＜5 were showed in Table 5 (now is Table 6)

Submitted filename: rebuttal letter.docx

Click here for additional data file.

8 Apr 2020

Validating the Chinese geriatric trigger tool and analyzing adverse drug event associated risk factors in elderly Chinese patients: A retrospective review

PONE-D-19-21999R1

Dear Dr. Xu,

We are pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it complies with all outstanding technical requirements.

Within one week, you will receive an e-mail containing information on the amendments required prior to publication. When all required modifications have been addressed, you will receive a formal acceptance letter and your manuscript will proceed to our production department and be scheduled for publication.

Shortly after the formal acceptance letter is sent, an invoice for payment will follow. To ensure an efficient production and billing process, please log into Editorial Manager at https://www.editorialmanager.com/pone/, click the "Update My Information" link at the top of the page, and update your user information. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org.

If your institution or institutions have a press office, please notify them about your upcoming paper to enable them to help maximize its impact. If they will be preparing press materials for this manuscript, you must inform our press team as soon as possible and no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

With kind regards,

Katie MacLure, PhD, MSc (dist)., BSc (1st)

Academic Editor

PLOS ONE

Additional Editor Comments (optional):

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. If the authors have adequately addressed your comments raised in a previous round of review and you feel that this manuscript is now acceptable for publication, you may indicate that here to bypass the “Comments to the Author” section, enter your conflict of interest statement in the “Confidential to Editor” section, and submit your "Accept" recommendation.

Reviewer #2: All comments have been addressed

Reviewer #3: All comments have been addressed

**********

2. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #2: Yes

Reviewer #3: Yes

**********

3. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #2: Yes

Reviewer #3: Yes

**********

4. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #2: Yes

Reviewer #3: Yes

**********

5. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #2: Yes

Reviewer #3: Yes

**********

6. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #2: Although the authors did not answered all my major or minor questions point-by-point, I managed to find the revised parts by myself.

Reviewer #3: All comments have been addressed in the revised manuscript.

One suggestion: It is better to separate the organ/system items in Table 3, or they might get mixed up.

**********

7. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #2: No

Reviewer #3: No

14 Apr 2020

PONE-D-19-21999R1

Validating the Chinese geriatric trigger tool and analyzing adverse drug event associated risk factors in elderly Chinese patients: A retrospective review

Dear Dr. Xu:

I am pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department.

If your institution or institutions have a press office, please notify them about your upcoming paper at this point, to enable them to help maximize its impact. If they will be preparing press materials for this manuscript, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org.

For any other questions or concerns, please email plosone@plos.org.

Thank you for submitting your work to PLOS ONE.

With kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Dr. Katie MacLure

Academic Editor

PLOS ONE