Evolution of Dementia Related to the Use of Alcohol in the French Nationwide Discharge Summary Database Between 2007 and 2017

Abstract

Background:

The French nationwide exhaustive hospital discharge database (PMSI) is used for activity-based payment of hospital services. We hypothesized that the release of articles about alcohol and dementia could influence the identification of these diagnoses in PMSI.

Methods:

We analyzed temporal evolution of coding for dementia and other persistent or late-onset cognitive impairment (OPLOCI) due to alcohol and other psychoactive substances in the PMSI database from 2007 to 2017 (285 748 938 inpatient stays). These codings use the International Classification of Diseases, 10th revision (ICD-10).

Results:

The number of inpatient stays with dementia and OPLOCI due to alcohol increased from 34 to 1704 from 2007 to 2017. While the number of diagnosed dementias remained stable at around 400 from 2013, the number of OPLOCIs increased 10-fold from 2013 to 2017. This increase was not found with dementia or OPLOCI due to other psychoactive substances than alcohol.

Conclusion:

Notoriety of a diagnosis in the literature seems to have an impact on the coding.

Introduction

Dementia is a clinical syndrome caused by neurodegeneration. It involves progressive decline in cognitive ability and capacity for independent living and functioning. ¹ The prevalence of dementia is estimated to be over 45 million people and could reach 115 million by 2050. ²

Among lifestyle-related factors for dementia, alcohol holds a special place. ³ Since 1997, low-to-moderate alcohol drinking has been proposed as a protective factor against the development of dementia. ^{4 -14} Contrasting findings exist, proposing low-to-moderate alcohol drinking as a risk factor ^{15 -17} or as an unassociated factor. ^{18 -26}

Chronic heavy alcohol drinking seems related to an increased risk of dementia. Heavy alcohol drinking may also be associated with other alcohol-related brain damage conditions (ethanol neurotoxicity, thiamine deficiency, Wernicke–Korsakoff syndrome, hepatic encephalopathy, vascular dementia). Heavy alcohol drinking is associated with other risk factors for dementia. ^{27 -29}

As intervention studies are not workable for alcohol exposure, the best evidence comes from retrospective cohort study and overview of epidemiological studies.

In early 2013, specific diagnosis of alcohol-related dementia has been discussed. ³⁰ In late 2015, a guidance produced by the National Institute for Health and Care Excellence (NICE) highlights that people aged 40 to 64 years who regularly drink alcohol have an increased risk of dementia. ^31,32

An association between early-onset dementia and alcohol use disorders has been found in the French nationwide exhaustive hospital discharge database (PMSI) between 2008 and 2013. ^29,33

The PMSI is used for activity-based payment of hospital services since 2004. The coding of the activity is the responsibility of the hospital services. In practice, residents and doctors code diagnoses and technical procedures; however, medical information technicians may code using hospitalization reports. Awareness of specific pathologies, via the media and scientific literature, could lead to changes in coding.

It has already been shown that notoriety affected spontaneous reports of event, particularly in the context of pharmacovigilance. ^34,35 We hypothesized that the medical coding in the databases is also subject to this notoriety bias and that the release of articles about alcohol and dementia could influence the identification of diagnoses of dementia related to the use of alcohol between 2007 and 2017.

Method

Study Design

This study consists of a retrospective cohort, based on secondary use of the PMSI database (presented later). We included all the inpatient stays between 2007 and 2017 from the website ScanSanté (a technical agency for hospital information).

Data Source

The PMSI database is the French nationwide exhaustive hospital discharge database. ³⁶ The database used in this study comprehends all the inpatient stays, from nonprofit and for-profit acute care hospitals (medicine, surgery, and obstetrics), excluding psychiatric hospitals and rehabilitation care centers. This database includes administrative data (admission and discharge dates and modes), demographic data (age, gender, geographic area), diagnoses encoded in International Classification of Diseases, 10th revision (ICD-10), ³⁷ medical procedure encoded in CCAM (French medical classification for clinical procedures), ³⁸ and other pieces of information. ³⁹ This information is anonymized and can be reused for research purposes. ³³ The database comprehends 285 748 938 inpatient stays over 11 years.

Inclusion Criteria

Dementia and related diseases encoding rules have been defined in 2006. ⁴⁰ Under these rules, we included inpatient stays having one of the following ICD-10 codes as principal diagnosis:

F1073: dementia due to alcohol

F1074: other persistent or late-onset cognitive impairment (OPLOCI) due to alcohol

The national coding recommendation for alcohol-induced dementia dates from March 2006. To ensure that the progression of alcohol-induced dementia was independent of dementia related to the use of other psychoactive substances, we have also analyzed the following codes:

F1173: dementia due to opioids

F1273: dementia due to cannabis

F1373: dementia due to sedatives and hypnotics

F1473: dementia due to cocaine

F1573: dementia due to stimulants

F1673: dementia due to hallucinogens

F1773: dementia due to tobacco

F1873: dementia due to volatile solvents

F1973: dementia due to multiple drugs

For each code, we also included F1*74 for OPLOCI (eg, F1174: other persistent or late-onset cognitive impairment due to opioids).

In comparison, we noted the number of inpatient stays for Alzheimer’s disease or dementia (codings F00 and G30) and for organic mental disorders, including symptomatic disorders: dementia of Alzheimer’s disease, vascular dementia, dementia during other diseases classified elsewhere, unspecified dementia, organic amnesia syndrome, and uninduced delirium; other mental disorders; personality and organic behavior disorders; and organic mental disorder (codings from F00 to F09).

Statistical Analysis

We performed descriptive statistics and reported categorical variables using the number and percentage of cases.

Results

Dementia and OPLOCI Due to Alcohol

From 2007 to 2012, the number of annual inpatient stays with a principal diagnosis of dementia due to alcohol increased from 20 to 432 and then stagnated until 2017.

From 2013, the number of annual inpatient stays with a principal diagnosis of OPLOCI due to alcohol was multiplied by 10 in 4 years (from 120 to 1247), with a sudden increase after 2015 (Figure 1).

Figure 1.

Evolution of inpatient stays for dementia and OPLOCI due to alcohol between 2007 and 2017 in France. OPLOCI indicates other persistent or late-onset cognitive impairment.

Dementia Related to the Use of Other Psychoactive Substances

Between 2007 and 2017, the number of inpatient stays for dementia and OPLOCI due to the use of other psychoactive substances than alcohol remained stable in France, with a peak of 23 in 2014. Unlike alcohol-induced dementia, there was no increase in the number of inpatient stays after this peak (Table 1).

Table 1.

Number of Inpatient Stays for Dementia and OPLOCI Due to the Use of Psychoactive Substances in France.

Principal Diagnosis		2007	2008	2009	2010	2011	2012	2013	2014	2015	2016	2017
Alcohol	Dementia	20	95	194	265	337	432	423	413	478	463	457
Alcohol	OPLOCI	14	21	45	72	89	121	120	275	382	924	1247
	Subtotal	34	116	239	337	426	553	543	688	860	1387	1704
Opioids	Subtotal	0	0	1	1	4	2	1	4	1	1	0
Cannabis	Subtotal	0	0	0	0	1	0	1	2	0	0	1
Sedatives and hypnotics	Subtotal	1	5	1	3	4	9	7	7	5	1	5
Cocaine	Subtotal	0	0	0	0	0	1	0	0	0	1	0
Stimulants	Subtotal	0	0	0	0	0	1	2	2	0	0	0
Hallucinogens	Subtotal	0	0	0	1	1	0	0	2	0	1	0
Tobacco	Subtotal	0	0	0	1	2	0	0	1	0	0	0
Volatile solvents	Subtotal	0	0	0	0	0	0	0	0	0	0	0
Multiple drugs	Subtotal	0	1	1	0	4	2	1	5	3	2	5
Alzheimer’s disease or dementia		46 313	49 108	45 706	41 792	4137	37 928	3844	31 097	29 235	28 716	25 856
Organic mental disorders		70 145	75 685	76 021	78 464	82 176	83 417	83 728	83 079	84 193	87 628	86 421
Inpatient stays (total)		23 339 533	23 781 314	24 297 929	24 709 773	25 259 627	25 828 027	26 476 016	27 087 492	27 600 941	28 538 316	28 829 970

Abbreviation: OPLOCI, other persistent or late-onset cognitive impairment.

Discussion

We observed a 3-fold increase in the number of inpatient stays with a principal diagnosis of dementia or OPLOCI due to alcohol from 2013 to 2017. This increase was not found with dementia or OPLOCI due to other psychoactive substances than alcohol.

Several hypotheses could explain our results. First, hospitalizations of patients with dementia may have increased over the 11-year period. It seems unlikely that in a study with a 15-year follow-up, the annual rate of hospitalization of patient with dementia was stable and represented 16.3 per 100 person years (95% CI 15.0-17.7). ⁴¹ Second, dementia or OPLOCI due to alcohol may have increased. Again, it seems unlikely that the number has increased by a factor of 50 in a few recent years. Assuming that the prevalence of dementia and related diseases is globally stable from 2008 to 2017, the observed variations may be due to a change in coding practices. ⁴² Thus, this trend may be due to the release of international publications about alcohol-related dementia. ^{3,11 -14,17,27,30,28} After the NHS communication and guidance in 2015 and the communication of a French team in 2016, the total number of dementia and OPLOCI due to alcohol coded in PMSI increased from 860 to 1704 in 2 years. ^31,33 Notoriety of a diagnosis in the literature seems to have an impact on the use of its coding, as we have previously suggested concerning vascular dementia. ⁴³

The principal strength of our study is to study all inpatient stays in France over 11 consecutive years, for a total of 285 748 938 inpatient stays.

Our study has potential limitations. The ICD-10 codes for inpatient stays were registered for billing purposes, and we did not confirm them versus clinical records. The use of the PMSI for activity-based payment of hospital services can lead to undercoding diseases that have no financial impact. Since 2012, government agency in charge of the PMSI database has been working on the Ovalide tool, a PMSI data validation tool, which will make it possible to identify low quality in PMSI data. The use of hospital coding to define the diagnosis is a recognized way of counting the cases of dementia. ⁴⁴ Our results suggest that coding recommendation about alcohol-related dementia is not respected because it is probable that alcoholic dementia is much larger, moreover in France. ⁴⁵ We study here the number of inpatient stays; some patients may have had several stays. Besides, we were interested in the evolution of coding, and there is no reason to believe that the annual number of stays of patients with alcohol-related dementia may have changed between 2007 and 2017.

Defining and identifying dementias in the major French administrative databases is an issue of interest: a recent review proposed several algorithms, 2 of which have been validated. ⁴⁶ Furthermore, dementia diagnosis requires fulfilling diagnostic criteria, such criteria exist for Alzheimer’s disease, ^47,48 frontotemporal dementia, ⁴⁹ vascular dementia, ^50,51 and dementia with Lewy bodies. ^52,53 There is no specific recommendation concerning alcohol-related dementias by the French Haute autorité de santé (equivalent to the British NICE) and no clear criteria for alcohol-related dementia. ³⁰

Conclusion

The findings of this study raise several issues: about care—we note the discrete recent emergence of alcohol-related dementia diagnosis and about research—notoriety of a diagnosis in the literature seems to have an impact on the use of its coding. Thus, it is important to keep in mind the notoriety of the diagnosis studied at the time of its analysis.