The Accuracy of Japanese Administrative Data in Identifying Acute Exacerbation of Idiopathic Pulmonary Fibrosis

ABSTRACT

BACKGROUND

This study aimed to develop criteria for identifying patients with acute exacerbation of idiopathic pulmonary fibrosis (AE-IPF) from Japanese administrative data and validate the pre-existing criteria.

METHODS

This retrospective, multi-center validation study was conducted at eight institutes in Japan to verify the diagnostic accuracy of the disease name for AE-IPF. We used the Japanese Diagnosis Procedure Combination data to identify patients with a disease name that could meet the diagnostic criteria for AE-IPF, who were admitted to the eight institutes from January 2016 to February 2019. As a reference standard, two respiratory physicians performed a chart review to determine whether the patients had a disease that met the diagnostic criteria for AE-IPF. Furthermore, two radiologists interpreted the chest computed tomography findings of cases considered AE-IPF and confirmed the diagnosis. We calculated the positive predictive value (PPV) for each disease name and its combination.

RESULTS

We included 830 patients; among them, 216 were diagnosed with AE-IPF through the chart review. We combined the groups of disease names and yielded two criteria: the criteria with a high PPV (0.72 [95% confidence interval 0.62 to 0.81]) and that with a slightly less PPV (0.61 [0.53 to 0.68]) but more true positives. Pre-existing criteria showed a PPV of 0.40 (0.31 to 0.49).

CONCLUSION

The criteria derived in this study for identifying AE-IPF from Japanese administrative data show a fair PPV. Although these criteria should be carefully interpreted according to the target population, our findings could be utilized in future database studies on AE-IPF.

BACKGROUND

Idiopathic pulmonary fibrosis (IPF) is a rare disease with a poor prognosis that is characterized by chronic progressive lung fibrosis. Although IPF is rare, it has a worse prognosis than numerous cancer types [1]. Moreover, hospitalization and mortality rates attributable to IPF have increased over recent years, which is suggestive of an increasing burden of this disease on healthcare services worldwide [2–4]. During the disease course, acute exacerbation (AE) may occur at an annual frequency of approximately 5–15%; further, the mortality rate is reported to be approximately 50–60% [5, 6]. Few randomized controlled trials (RCTs) with limited external validity have investigated AE-IPF [7, 8]. Therefore, there is a need for large observational studies on a wider patient population using real-world data to complement RCTs [9].

However, there have been rare seen large-scale database studies on AE-IPF. This could be attributed to the difficulty in identifying cases. In large-scale database research, case identification is essential, with validation studies being crucial for this purpose. The REporting of studies Conducted using Observational Routinely collected health Data (RECORD) statement mentions that “Any validation studies of the codes or algorithms used to select the population should be referenced” [10]. Regarding AE-IPF, one validation study has been performed in the US [11]; however, the positive predictive value (PPV) was not very high (0.621 [95% confidence interval [CI] 0.533 to 0.704]). Additionally, it remains unclear whether the data can be applied in other countries.

This study aimed to develop criteria for identifying patients with AE-IPF, as well as validate the criteria in the previous study, using multi-center chart data in Japan.

METHODS

We performed and reported this study in accordance with the reporting guidelines for assessing the quality of validation studies of health administrative data (Table S1) [12].

SETTINGS AND PATIENTS

This retrospective, multicenter, validation study was conducted to verify the diagnostic accuracy of the disease name for AE-IPF. We performed study in eight Japanese tertiary hospitals: Saiseikai Kumamoto Hospital, Hyogo Prefectural Amagasaki General Medical Center, Iizuka Hospital, Kobe City Medical Center General Hospital, Kameda Medical Center, Hoshigaoka Medical Center, Okinawa Chubu Hospital, and the Japanese Red Cross Medical Center.

We included patients aged ≥40 years with a disease name that could meet the diagnostic criteria for AE-IPF (Table 1), who were admitted to the aforementioned eight institutes from January 2016 to February 2019. These criteria were developed in advance by two respiratory specialists. We excluded patients with disease names for secondary interstitial pneumonia, including collagen vascular disease-associated interstitial lung disease and chronic and hypersensitivity pneumonitis, as well as those for malignant neoplasms (Table 2). We did not exclude patients with disease names for malignant neoplasms with the suffixes “post-operation” or “post-radiotherapy”. We ignored all “suspected” disease names.

Table 1 Disease names adopted as inclusion criteria

ICD-10 code 1	disease name 1*	suffix	ICD-10 code 2	disease name 2**	suffix	emergency admission
J841	idiopathic pulmonary fibrosis	AE	—	—	—	—
J841	idiopathic interstitial pneumonia	AE	—	—	—	—
J841	usual interstitial pneumonia	AE	—	—	—	—
J841	pulmonary fibrosis	AE	—	—	—	—
J841	diffuse interstitial pneumonia	AE	—	—	—	—
J849	interstitial pneumonia	AE	—	—	—	—
J841	acute interstitial pneumonia	—	—	—	—	—
J841	diffuse interstitial pneumonia	—	—	—	—	—
J841	diffuse alveolar damage	—	—	—	—	—
J80	acute respiratory distress syndrome	—	—	—	—	—
J984	acute lung injury	—	—	—	—	—
J702	acute drug—induced interstitial lung disorders	—	—	—	—	—
J704	drug-induced interstitial pneumonia	—	—	—	—	—
J10～J111 J15～189 J690	pneumonia, Influenza, aspiration pneumonia	—	J841	idiopathic pulmonary fibrosis	AE	—
		—	J841	idiopathic interstitial pneumonia	AE	—
		—	J841	usual interstitial pneumonia	AE	—
		—	J841	pulmonary fibrosis	AE	—
		—	J841	diffuse interstitial pneumonia	AE	—
		—	J841	interstitial pneumonia	AE	—
		—	J841	idiopathic pulmonary fibrosis	—	—
		—	J841	idiopathic interstitial pneumonia	—	—
		—	J841	usual interstitial pneumonia	—	—
		—	J841	pulmonary fibrosis	—	—
		—	J841	diffuse interstitial pneumonia	—	—
		—	J841	interstitial pneumonia	—	—
J841	idiopathic pulmonary fibrosis	—	—	—	—	○
J841	idiopathic interstitial pneumonia	—	—	—	—	○
J841	usual interstitial pneumonia	—	—	—	—	○
J841	pulmonary fibrosis	—	—	—	—	○
J841	diffuse interstitial pneumonia	—	—	—	—	○
J849	interstitial pneumonia	—	—	—	—	○

AE, acute exacerbation; DPC, diagnosis procedure combination; ICD-10, the International Statistical Classification of Diseases and Related Health Problems, Tenth Revision

* Disease name in any categories in the DPC data

** Disease name in co-morbidities present at admission and conditions occurring after admission

Table 2 Disease names adopted as exclusion criteria

ICD-10 code	disease name*
C00～C96	malignant neoplasms**
D860	sarcoidosis of lung
J60	coal worker’s pneumoconiosis
J61	asbestosis
J64	pneumoconiosis
J67	hypersensitivity pneumonitis due to organic dust
J990·M0510	rheumatoid arthritis-associated interstitial lung disease
J991·M351	collagen vascular disease-associated interstitial lung disease
J991·M330	juvenile dermatomyositis-associated interstitial lung disease
J991·M321	systemic lupus erythematosus-associated interstitial lung disease
J991·M332	polymyositis-associated interstitial lung disease
J991·M331	dermatomyositis-associated interstitial lung disease
J991·M313	lung involvement in granulomatosis with polyangiitis
M348	lung involvement in systemic sclerosis

* Disease name in any category in the DPC data

** not exclude if they have attached suffix of “post-operation” or “post-radiotherapy”.

Regarding disease names, we extracted data from the Diagnosis Procedure Combination (DPC), which was launched at Japanese acute-care hospitals across the country in 2003 [13, 14]. The following data were extracted from the DPC: some clinical information (e.g. age, gender, etc.), admission route (e.g. scheduled or emergency admission), diagnoses, surgeries and procedures, medications, outcomes, etc. The patients were recorded with diagnostic names using the International Statistical Classification of Diseases and Related Health Problems, Tenth Revision (ICD-10) codes. The diagnoses included the following six categories: main diagnosis, diagnosis that triggered hospitalization, diagnosis using the most medical resources, diagnosis using the second most medical resources, comorbidities present at admission, and conditions occurring after admission. One diagnostic name is coded for each main diagnosis, diagnosis that triggered hospitalization, diagnosis using the most medical resources, and diagnosis using the second most medical resources. Meanwhile, up to 10 diagnostic names can be coded for comorbidities present at admission and conditions occurring after admission. Furthermore, in the Japanese DPC data, modifiers can be added to the disease name in ICD-10 (e.g., regions such as left and right, recurrence, acute exacerbation, suspected, etc.).

This study was approved by the Ethics Committee of leading institute of this study (Kyoto University Graduate School and Faculty of Medicine) (Approved Number: R2071) and other committed institutes. The requirement for written informed consent from the participants was waived because of the retrospective design. Meanwhile, we offered the patients an opportunity to opt-out.

REFERENCE STANDARD

We confirmed the AE-IPF diagnosis using the following two steps:

1) At least two respiratory physicians at each institute reviewed the medical records and chest computed tomography (CT) findings on admission. Subsequently, they determined whether the patients met the following diagnostic criteria: an official ATS/ERS/JRS/ALAT clinical practice guideline for IPF diagnosis [15] and the international working group diagnostic criteria for AE-IPF [16]. An IPF diagnosis was determined if UIP or probable UIP pattern was observed on HRCT findings, even without surgical lung biopsy (SLB). If SLB was performed, IPF was considered based on specific combinations of HRCT and histopathological patterns following the ATS/ERS/JRS/ALAT clinical practice guidelines [15]. Discrepancies were resolved through discussion or consultation with a third respiratory physician, as appropriate.

2) Two expert chest radiologists interpreted chest CT findings of cases considered to be AE-IPF in 1) and determined whether the patients met the aforementioned two diagnostic criteria. If both radiologists considered that a case was not AE-IPF, it was determined as not AE-IPF. On the other hand, if a case was considered as not as AE-IPF by only one radiologist, the case was determined as AE-IPF.

Multidisciplinary discussion (MDD), which is held for diagnostic decision-making regarding patients with interstitial lung disease who are clinically suspected to have IPF, has been described in the official ATS/ERS/JRS/ALAT clinical practice guideline [15]. However, MDD is a conditional recommendation that is not widely held in Japanese hospitals. Therefore, we decided not to consider MDD as a diagnostic requirement.

DEVELOPING CRITERIA

First, we assessed the diagnostic accuracy of each disease name in the inclusion criteria, with disease names showing high diagnostic accuracy being adopted without modification. Next, we explored combinations of disease names with high diagnostic accuracy for disease names with a high number of true positives. We sought to create two criteria: 1) the criteria with the higher PPV (narrow criteria) and 2) criteria with slightly less PPV but more true positives (broad criteria). Based on previous studies [11, 17, 18], we excluded disease names of secondary interstitial lung diseases (ILDs) and included those of acute respiratory failure (ARF) (Table 3). Further, we excluded disease names with pneumonia, influenza, and aspiration pneumonia and addition of that of IPF.

Table 3 ICD-10 codes for interstitial lung diseases (ILDs) other than idiopathic pulmonary fibrosis (IPF) and acute respiratory failure.

ILDs other than IPF
ICD-10 code	disease name
C966	Langerhans-cell histiocytosis*
D869	Sarcoidosis
E756	Lipidoses
J998·E831	Idiopathic pulmonary hemosiderosis
E85	Amyloidosis
J60	coal worker’s pneumoconiosis*
J61	asbestosis*
J62	silicosis or talcosis
J63	berylliosis and other inorganic dusts
J64	unspecified pneumoconiosis*
J67	hypersensitivity pneumonitis*
J82	Pulmonary eosinophilia, not elsewhere classified
J82	Extrinsic allergic alveolitis
J684	Chronic respiratory conditions due to chemicals, gases, fumes, and vapors
J701	Chronic and other pulmonary manifestations due to radiation
J708	Respiratory conditions due to other specified external agents
J840	Alveolar and parieto-alveolar conditions
J841	Postinflammatory pulmonary fibrosis
J99	Respiratory disorders in diseases classified elsewhere
K50	Regional enteritis
M05, M06	Rheumatoid arthritis
M310	Goodpasture’s syndrome
M313	Wegener’s granulomatosis
M317	Microscopic polyangiitis
M318	Other specified necrotizing vasculopathies
M321, M329	Systemic lupus erythematosus
M330, M331, M339	Dermatomyositis
M332	Polymyositis
M340, M341, M349	Systemic sclerosis
M348	Lung involvement in systemic sclerosis*
M350	Sjogren’s disease
M351	other overlap syndrome
M45	Ankylosing spondylitis
Q850	Neurofibromatosis
Q851	Tuberous sclerosis
Acute respiratory failure
ICD-10 code	disease name
J960	acute respiratory failure

* We already excluded these disease names as exclusion criteria

EXTERNAL VALIDATION OF THE CRITERIA OF A PREVIOUS STUDY

To compare the PPV with the aforementioned newly developed criteria, we externally validated the criteria of a previous study in the US [11] (pre-existing criteria) using DPC data extracted from the aforementioned eight hospitals. The inclusion criteria were as follows: 1) age ≥50 years; 2) having at least one claim to a specific diagnostic code for IPF (either ICD10 code J84.9, J84.10, J84.111, J84.112, or ICD-9 correspond); 3) having acute respiratory symptoms or ARF (J84.114, J84.9, J96.00-02, J96.20-22, J96.91-92, or ICD-9 correspond) and subsequent hospital admission; and 4) lacking secondary ILDs.[11] In accordance with the Japanese DPC data used in this study, we modified this criteria as follows: 1) age ≥50 years; 2) having at least one following disease name (either ICD10 code J84.9 [interstitial pneumonia (IP)] or J84.1 [IPF or idiopathic interstitial pneumonia (IIP) or pulmonary fibrosis] in three diagnosis categories (main diagnosis, diagnosis that triggered hospitalization, diagnosis using the most medical resources); 3) having at least one following disease names (ICD10 code J84.1 [acute interstitial pneumonia] or J96.0 [ARF]) in any diagnosis category, and emergency admission; and 4) lacking secondary ILDs.

STATISTICAL ANALYSIS

Interobserver agreement for the AE-IPF diagnosis between both respiratory specialists was evaluated using kappa statistics. For cases diagnosed as AE-IPF by two respiratory physicians, the interobserver agreement for AE-IPF diagnosis between both radiologists was assessed using the AC₁-index since most cases were considered to be true positive cases [19]. We calculated the PPV as true positives / (true positives + false positives) for every disease name and their combinations. Further, we calculated the 95% confidence intervals (CIs) for binomial distribution using the exact method. We could not calculate the sensitivity, specificity, and negative predictive values since we did not perform a chart review of patients without target ICD-10 codes. Validation studies for only obtaining a PPV have reported that a sample size of approximately 100 is sufficient [20], because with that number of samples, the 95% CI falls within the range of the point estimate ± 0.1. We sought to identify patients from various disease names and search for algorithms that would improve PPV. Therefore, regarding the narrow criteria, we identified the final criteria with the largest PPV out of approximately 100 cases. Statistical analyses were performed using STATA/SE version.16.0 (Stata Corp., College Station, TX, USA).

RESULTS

This study included 822 patients; among them, 212 were diagnosed with AE-IPF by chart review. There was substantial interobserver agreement for AE-IPF diagnosis (kappa = 0.80 [0.75 to 0.85]). Among the patients who were considered to have AE-IPF by both respiratory physicians, none were considered not to have AE-IPF by both radiologists; moreover, there was excellent interobserver agreement between both radiologists (AC₁-index = 0.99 [0.98 to 1]).

Table 4 and Table S2 show the PPV and 95% CI for each disease name. Although the disease name of AE-IPF and acute exacerbation of IIP (AE-IIP) had the highest and second-highest PPV, respectively, the number of patients was small. The PPV for all the other disease names was low.

Table 4 The positive predictive value (PPV) and 95% confidence intervals (CI) of patients diagnosed with acute exacerbation of idiopathic pulmonary fibrosis within each algorithm

disease names	total (n)	True Positives	PPV	95% CI
IPF + AE	5	4	0.8	0.28 to 0.99
IIP + AE	32	24	0.75	0.57 to 0.89
(IIP + AE) − secondary ILDs*	30	23	0.77	0.58 to 0.90
(IIP + AE) − secondary ILDs + ARF*	22	17	0.77	0.55 to 0.92
(IIP + AE) − secondary ILDs − infection	29	22	0.76	0.55 to 0.92
(IIP + AE) − secondary ILDs − infection + ARF	21	16	0.76	0.56 to 0.90
(IIP + AE) + IPF	6	6	1	0.54 to 1
IP + AE	105	59	0.56	0.46 to 0.66
(IP + AE) − secondary ILDs	95	57	0.6	0.49 to 0.70
(IP + AE) − secondary ILDs + ARF	75	50	0.67	0.55 to 0.77
(IP + AE) − secondary ILDs − infection	77	48	0.62	0.51 to 0.73
(IP + AE) − secondary ILDs − infection + ARF	60	42	0.7	0.57 to 0.81
(IP + AE) + IPF	15	9	0.6	0.32 to 0.84
(IP + AE) + IPF − secondary ILDs − infection	12	8	0.67	0.35 to 0.90
IPF + emergency admission	40	16	0.4	0.25 to 0.57
IPF + emergency admission + ARF	15	10	0.67	0.38 to 0.88
IPF + emergency admission − secondary ILDs − infection + ARF	12	8	0.67	0.35 to 0.90
IPF + emergency admission + (IP + AE)	4	2	0.5	0.068 to 0.93
IIP + emergency admission	152	35	0.23	0.17 to 0.31
IIP + emergency admission + ARF	47	19	0.40	0.26 to 0.56
IIP + emergency admission − secondary ILDs − infection + ARF	21	7	0.33	0.15 to 0.57
IIP + emergency admission + (IP + AE)	3	2	0.67	0.094 to 0.99
IP + emergency admission	140	38	0.27	0.20 to 0.35
IP + emergency admission + ARF	78	27	0.35	0.24 to 0.46
IP + emergency admission − secondary ILDs − infection + ARF	45	14	0.31	0.18 to 0.47
final algorithm
Narrow criteria
(IPF + AE) OR ((IIP + AE) − secondary ILDs − infection) OR ((IIP + AE) + IPF) OR ((IP + AE) − secondary ILDs − infection + ARF)	94	68	0.72	0.62 to 0.81
Broad criteria
(IPF + AE) OR (IIP + AE) OR ((IP + AE) − secondary ILDs + ARF) OR (IPF + emergency admission + (ARF OR (IP + AE))) OR (IIP + emergency admission + (ARF OR (IP + AE)))	168	102	0.61	0.53 to 0.68
Pre-existing criteria
((IPF OR IIP OR pulmonary fibrosis OR IP) + (AIP OR ARF) + emergency admission − secondary ILDs)	126	50	0.40	0.31 to 0.49

AE, acute exacerbation; AIP, acute interstitial pneumonia; ARF, acute respiratory failure; IIP, idiopathic interstitial pneumonia; ILDs, interstitial lung diseases; IP, interstitial pneumonia; IPF, idiopathic pulmonary fibrosis

“+” represent AND; “−“ represent NOT

* disease names of secondary ILDs and ARF are described in Table 3

NARROW CRITERIA

As a result of examining each of the disease names, we found that the following five disease names (AE-IIP, acute exacerbation of interstitial pneumonia, AE-IP, IPF + emergency admission, IIP + emergency admission, and IP + emergency admission) in three categories (main diagnosis, diagnosis that triggered hospitalization, or diagnosis using the most medical resources) contained many true positive patients. Therefore, we explored algorithms that would result in higher PPV for these disease names. We attempted to exclude disease names related to secondary ILDs (any of all categories) and infection (in four categories; main diagnosis, diagnosis that triggered hospitalization, diagnosis using the most medical resources, or comorbidities present at admission), and add to disease names for ARF (any of all categories) and IPF (any of all categories). Consequently, we observed an increased PPV for several combinations of disease names.

Finally, we combined the groups of disease names with a higher PPV and yielded a criteria (AE-IPF OR (AE-IIP − secondary ILDs − infection) OR (AE-IIP + IPF) OR (AE-IP − secondary ILDs − infection + ARF) with a PPV of 0.72 (95% CI 0.62 to 0.81) (Table 4).

BROAD CRITERIA

Using the five disease names, we searched for a criteria that would include more true-positive patients even with a slightly inferior PPV. Finally, we created a criteria (AE-IPF OR AE-IIP OR [AE-IP − secondary ILDs + ARF] OR [IPF + emergency admission + [ARF OR AE-IP] OR [IIP + emergency admission + (ARF OR AE-IP)] with a PPV of 0.61 (95% CI 0.53 to 0.68) (Table 4).

PRE-EXISTING CRITERIA

The validation results for the pre-existing criteria (IPF OR IIP OR pulmonary fibrosis OR IP) + (AIP OR ARF) + emergency admissions − secondary ILDs) showed a PPV of 0.40 (95% CI 0.31 to 0.49) (Table 4).

DISCUSSION

Using Japanese administrative data, we created criteria for identifying patients with AE-IPF and validated the pre-existing criteria. It was difficult to accurately identify true positive AE-IPF cases only based on a single disease name. Consequently, we excluded or added various other disease names and emergency admissions, which improved the AE-IPF diagnosis accuracy. The pre-existing criteria had a relatively low PPV. The PPV for the final criteria were as follows: narrow criteria, 0.72 (0.62 to 0.81); broad criteria, 0.61 (0.53 to 0.68); and pre-existing criteria, 0.40 (0.31 to 0.49).

The diagnostic accuracy of the narrow and broad criteria showed a fair PPV. Currently, the only validation study on AE-IPF derived a criteria with a PPV of 0.621 (95% CI 0.533 to 0.704) [11], which was lower and comparable to the PPV of the narrow and broad criteria, respectively. Most disease name combinations in the narrow criteria used the suffix acute exacerbation, which is unique to the DPC data and may have attributed to the higher PPV compared with that in the previous study. Meanwhile, the validation results for the pre-existing criteria revealed a relatively low PPV (0.40 [95% CI 0.31 to 0.49]). This indicates that it may be better to use the narrow or broad criteria rather than the pre-existing criteria in the Japanese DPC data.

Meanwhile, our findings are suggestive of the difficulty of AE-IPF diagnosis only based on administrative data. As shown in Table 4, there were only four true-positive patients with a disease name of AE-IPF. IPF is a rare disease; moreover, its diagnosis is heavily dependent on high-resolution CT (HRCT). Additionally, IPF diagnosis is sometimes difficult since it requires the exclusion of all other diseases that cause chronic fibrosis of the lungs. Furthermore, in case of acute respiratory worsening in fibrotic ILDs [21], it can be challenging for clinicians to distinguish between AE and others. Therefore, it is very difficult to accurately diagnose AE-IPF; moreover, many true positive patients could have been included among those with disease names of IIP or IP as well as IPF with or without AE. To exclude secondary ILDs, we referred to the ICD-9 and 10 codes used in previous IPF database studies [11, 17, 18]. For AE, we referred to a previous study and added a code for ARF [11]. Combining these codes with IIP and IP increased the PPV; therefore, these codes may be useful for excluding secondary ILDs and including AEs. For diseases that are rare and difficult to diagnose, a PPV of approximately ≥70% is considered reasonable [22, 23]. Therefore, given the aforementioned difficulties in diagnosing AE-IPF, the PPVs obtained in our study could be considered reasonable.

RESEARCH IMPLICATIONS

We believe our findings regarding the criteria for identifying patients with AE-IPF can contribute to database research in Japan. Meanwhile, using the narrow criteria, only about 1/3 of the cases were true positives. To detect more patients with AE-IPF, it may be better to use the broad criteria even though it has a lower PPV. Additionally, the narrow and broad criteria are both insufficient for selecting patients with AE-IPF patients without omission. It should be recognized that some patients with disease names, including acute respiratory distress syndrome or drug-induced interstitial pneumonia, may meet the criteria for AE-IPF. Furthermore, the institutions involved in this study were tertiary hospitals that treat numerous patients with AE-IPF. Although DPC data are collected from numerous acute care hospitals, there were not all proficient in treating patients with AE-IPF; thus, our findings may not be generalizable as representative data for Japan.

We created criteria for defining AE-IPF with not very high PPV. Although the present study and a previous validation study were conducted only using disease names and hospitalization information, the DPC data have rich data on medications and procedures. Therefore, by adding these information, such as information on administration of systemic corticosteroid therapy, it may be possible to employ an algorithm with a higher PPV. Additionally, using PPV alone cannot allow for a database study that covers the entire AE-IPF population. Therefore, there is a need for studies calculating the sensitivity and specificity using methods, including random sampling or “all possible cases” sampling [24].

STRENGTH AND LIMITATIONS

Our study has several strengths. First, this is the first validation study on AE-IPF using Japanese DPC data. Second, this was a multi-center study with a sufficient sample size. Third, AE-IPF diagnosis was more reliable because of the interpretation by two chest radiologists in addition to at least two pulmonologists.

This study had several limitations. First, since this study was conducted at tertiary emergency hospitals with respiratory specialists, it remains unclear whether it can be adapted to the entire DPC data in Japan. Therefore, our findings should be externally validated in other facilities in Japan. Second, this was a validation study of the Japanese DPC data, which is unique to Japan and does not apply to other countries. Finally, we cannot eliminate the possibility of misclassification of the IPF diagnosis (e.g., hypersensitivity pneumonitis and collagen vascular disease) or AE. However, we believe that this was reduced by the interpretation of the chest CT findings by two expert chest radiologists.

CONCLUSION

In conclusion, this study derived criteria for identifying AE-IPF from Japanese administrative data with a fair PPV. When studying AE-IPF in Japanese administrative data using the narrow or broad criteria, researchers should carefully interpret the target population.

FUNDING

The English editing fee was supported by Saiseikai Kumamoto Hospital. The funder did not play any role in the design of the study and collection, analysis and interpretation of data

CONFLICTS OF INTEREST

Keisuke Anan received a research grant from Systematic Review Workshop Peer Support Group and Fujiwara Memorial Foundation for other research purposes. Yuki Kataoka received a research grant from the Systematic Review Workshop Peer Support Group, the Japan Osteoporosis Foundation, and Yasuda Memorial Medical Foundation for other research purposes.

ACKNOWLEDGMENTS

We are grateful to Asuka Iwashita (Saiseikai Kumamoto Hospital), Hiromi Ide (Iizuka Hospital), Keiko Sakuragawa (Kobe City Medical Center General Hospital), Kyoko Wasai (Hyogo Prefectural Amagasaki General Medical Center), Hiroshi Harada (Kameda Medical Center), Megumi Kunita (Hoshigaoka Medical Center), Daiki Chinen (Okinawa Chubu Hospital), and Misao Ebihara (Japanese Red Cross Medical Center) for retrieving data. We would like to thank Editage (www.editage.com) for English language editing.

AUTHOR CONTRIBUTIONS

KA is the main author and contributed to the design of the project, data collection and analysis, and wrote the draft of the manuscript. YK contributed to the design of the project and data collection and analysis. KI, KK, TJ, KF, KT, RT, HI, TN, TK, and MI contributed data collection. YY contributed to the design of the project and supervised the study. All authors contributed to data interpretation, revised the draft, and reviewed and approved the final manuscript.

REFERENCES

• 1.   du Bois  RM. An earlier and more confident diagnosis of idiopathic pulmonary fibrosis. Eur Respir Rev 2012;21:141–146.
• 2.   Hutchinson  J,  Fogarty  A,  Hubbard  R,  McKeever  T. Global incidence and mortality of idiopathic pulmonary fibrosis: A systematic review. Eur Respir J 2015;46:795–806.
• 3.   Hutchinson  JP,  McKeever  TM,  Fogarty  AW,  Navaratnam  V,  Hubbard  RB. Increasing global mortality from idiopathic pulmonary fibrosis in the twenty-first century. Ann Am Thorac Soc 2014;11:1176–1185.
• 4.   Navaratnam  V,  Fogarty  AW,  Glendening  R,  McKeever  T,  Hubbard  RB. The increasing secondary care burden of idiopathic pulmonary fibrosis: hospital admission trends in England from 1998 to 2010. Chest 2013;143:1078–1084.
• 5.   Richeldi  L,  du Bois  RM,  Raghu  G,  Azuma  A,  Brown  KK,  Costabel  U, et al. Efficacy and safety of nintedanib in idiopathic pulmonary fibrosis. N Engl J Med 2014;370:2071–2082.
• 6.   Song  JW,  Hong  SB,  Lim  CM,  Koh  Y,  Kim  DS. Acute exacerbation of idiopathic pulmonary fibrosis: incidence, risk factors and outcome. Eur Respir J 2011;37:356–363.
• 7.   Raghu  G,  Collard  HR,  Egan  JJ,  Martinez  FJ,  Behr  J,  Brown  KK, et al. An official ATS/ERS/JRS/ALAT statement: Idiopathic pulmonary fibrosis: Evidence-based guidelines for diagnosis and management. Am J Respir Crit Care Med 2011;183:788–824.
• 8.   Homma  S,  Bando  M,  Azuma  A,  Sakamoto  S,  Sugino  K,  Ishii  Y, et al. Japanese guideline for the treatment of idiopathic pulmonary fibrosis. Respir Investig 2018;56:268–291.
• 9.   Schneeweiss  S,  Avorn  J. A review of uses of health care utilization databases for epidemiologic research on therapeutics. J Clin Epidemiol 2005;58:323–337.
• 10.   Benchimol  EI,  Smeeth  L,  Guttmann  A,  Harron  K,  Moher  D,  Petersen  I, et al. The REporting of studies Conducted using Observational Routinely-collected health Data (RECORD) statement. PLoS Med 2015;12:e1001885.
• 11.   Farrand  E,  Vittinghoff  E,  Ley  B,  Butte  AJ,  Collard  HR. Corticosteroid use is not associated with improved outcomes in acute exacerbation of IPF. Respirology 2020;25:629–635.
• 12.   Benchimol  EI,  Manuel  DG,  To  T,  Griffiths  AM,  Rabeneck  L,  Guttmann  A. Development and use of reporting guidelines for assessing the quality of validation studies of health administrative data. J Clin Epidemiol 2011;64:821–829.
• 13.   Okamura  S,  Kobayashi  R,  Sakamaki  T. Case-mix payment in Japanese medical care. Health Policy 2005;74:282–286.
• 14.   Yasunaga  H,  Ide  H,  Imamura  T,  Ohe  K. Impact of the Japanese Diagnosis Procedure Combination-based Payment System on cardiovascular medicine-related costs. Int Heart J 2005;46:855–866.
• 15.   Raghu  G,  Jardin  MR,  Myers  JL,  Richeldi  L,  Ryerson  CL,  Lederer  DJ, et al. Diagnosis of idiopathic pulmonary fibrosis. An Official ATS/ERS/JRS/ALAT Clinical Practice Guideline. Am J Respir Crit Care Med 2018;198:e44–e68.
• 16.   Collard  HR,  Ryerson  CJ,  Corte  TJ,  Jenkins  G,  Kondoh  Y,  Lederer  DJ, et al. Acute exacerbation of idiopathic pulmonary fibrosis. An international working group report. Am J Respir Crit Care Med 2016;194:265–275.
• 17.   Raghu  G,  Weycker  D,  Edelsberg  J,  Bradford  WZ,  Oster  G. Incidence and prevalence of idiopathic pulmonary fibrosis. Am J Respir Crit Care Med 2006;174:810–816.
• 18.   Lai  CC,  Wang  CY,  Lu  HM,  Chen  L,  Teng  NC,  Yan  YH, et al. Idiopathic pulmonary fibrosis in Taiwan—A population-based study. Respir Med 2012;106:1566–1574.
• 19.   Shankar  V,  Bangdiwala  SI. Observer agreement paradoxes in 2 × 2 tables: Comparison of agreement measures. BMC Med Res Methodol 2014;14:100.
• 20.   Cutrona  SL,  Toh  S,  Iyer  A,  Foy  S,  Cavagnaro  E,  Forrow  S, et al. Design for validation of acute myocardial infarction cases in Mini-Sentinel. Pharmacoepidemiol Drug Saf 2012;21:274–281.
• 21.   Moua  T,  Westerly  BD,  Dulohery  MM,  Daniels  CE,  Ryu  JH,  Lim  KG. Patients with fibrotic interstitial lung disease hospitalized for acute respiratory worsening: A large cohort analysis. Chest 2016;149:1205–1214.
• 22.   Segal  JB,  Powe  NR. Accuracy of identification of patients with immune thrombocytopenic purpura through administrative records: A data validation study. Am J Hematol 2004;75:12–17.
• 23.   Wahl  PM,  Terrell  DR,  George  JN,  Rodgers  JK,  Uhl  L,  Cataland  S, et al. Validation of claims-based diagnostic codes for idiopathic thrombotic thrombocytopenic purpura in a commercially-insured population. Thromb Haemost 2010;103:1203–1209.
• 24.   Widdifield  J,  Ivers  NM,  Young  J,  Green  D,  Jaakkimainen  L,  Butt  DA, et al. Development and validation of an administrative data algorithm to estimate the disease burden and epidemiology of multiple sclerosis in Ontario, Canada. Mult Scler 2015;21:1045–1054.