National U.S. time-trends in opioid use disorder hospitalizations and associated healthcare utilization and mortality.

BACKGROUND: The opioid epidemic is a major public health crisis in the U.S. Contemporary data on opioid use disorder (OUD) related hospitalizations are needed. Our objective was to assess whether OUD hospitalizations and associated mortality are increasing over time and examine the factors associated healthcare utilization and mortality. METHODS AND
FINDINGS: We examined the rates of OUD hospitalizations and associated mortality using the U.S. National Inpatient Sample (NIS) data from 1998-2016. Multivariable-adjusted logistic regression assessed the association of demographic, clinical and hospital characteristics with inpatient mortality and healthcare utilization (total hospital charges, discharge to a rehabilitation facility, length of hospital stay) during the index hospitalization for opioid use disorder. We calculated the odds ratio (OR) and 95% confidence intervals (CI). We estimated 781,767 OUD hospitalizations. The rate of OUD hospitalization and associated mortality (/100,000 overall NIS hospitalizations) increased from 59.8 and 1.2 in 1998-2000 to 190.7 and 5.9 in 2015-16, respectively. In the multivariable-adjusted analysis, the following factors were associated with worse outcomes; compared to age <34 years, older age was associated with higher risk of hospital charges above the median and length of stay >3 days, slightly higher risk of discharge to a rehabilitation facility. Higher Deyo-Charlson score was associated with higher hospital charges, length of hospital stay, and inpatient mortality. Women had lower odds of inpatient mortality than men and blacks had lower odds of mortality than whites.
CONCLUSIONS: Rising OUD hospitalizations from 1998 to 2016 and increasing associated inpatient mortality are concerning. Certain groups are at higher risk of poor utilization outcomes and inpatient mortality. Resources and healthcare policies need to focus on the high-risk group to reduce mortality and associated utilization.

Introduction

The opioid epidemic in the U.S. is a concern for providers, hospitals, policy-makers and the public [1-3]. The opioid epidemic is associated with significant mortality with calls for action to end the epidemic [4,5]. Based on national vital statistics data, the Centers for Disease Control (CDC) reported that 28,647 opioid-related deaths in 2014 that increased to 33,091 in 2015 (16% increase) [1], and to 42,249 deaths in 2016 (47% increase) [6]. The opioid overdose death rate increased from 2000 to 2014 [3] and continued the upward trend, increasing from 9.0 per 100,000 in 2014 to 10.4 in 2015 [1].

In addition to examining opioid use disorder (OUD)-related mortality, hospitalizations associated with OUD can help us better understand the opioid epidemic. A recent study of U.S. national inpatient sample (NIS) documented that nearly half a million hospitalizations yearly included a diagnosis of OUD (in any position, primary or secondary) [7]. Regional and demographic differences exist in prescription opioid and heroin-related overdose hospitalizations [8]. The Centers for Disease Control (CDC) issued a guideline to reduce the overutilization of prescription opioid use as a potential solution to the opioid epidemic [9]. Various state and federal agencies, including the drug enforcement agencies, have been monitoring narcotic prescription patterns [10].

Due to the limited data on hospitalizations for OUD without opioid overdose, detoxification or rehabilitation services, we aimed to examine hospitalizations related to this clinical problem. We examined time-trends in the OUD hospitalizations and the associated healthcare utilization and mortality, and assessed the factors associated with healthcare utilization and mortality during the OUD-associated hospitalizations.

Materials and methods

Data source

We included discharges from the Healthcare Cost and Utilization Project’s (HCUP) National Inpatient Sample (NIS) from 1998 to 2016. The NIS is a 20% stratified sample of hospital discharges, designed for creating national estimates of all hospitalizations in the U.S. The NIS changed design in 2012 from a 20% sample of hospitals to a 20% sample of discharges from hospitals. We used the recommended trend weights from the HCUP documentation to allow analyses across multiple years. The University of Alabama at Birmingham’s Institutional Review Board approved this study (X120207004) and waived the need for informed consent for this database study since these national data are de-identified. All investigations were conducted in conformity with ethical principles of research.

Study cohort

We identified hospitalizations for OUD based on the presence of any of the following International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) or Tenth Revision, Clinical Modification (ICD-10-CM) diagnostic codes for opioid dependence, abuse, or poisoning in the primary diagnosis position, as by the Agency for Healthcare Research and Quality [11]: ICD-9-CM: 304.0x, 304.7x, 305.5x, 965.0x, E850.0, E935.0, F111.xxx, F112.xxx, T40.1X1x-4x,T40.2X1x-4x, or T40.3X1x-4x. We excluded hospitalizations with ICD-9-CM diagnostic or procedure codes corresponding to drug/alcohol counseling or rehabilitation/detoxification including diagnostic codes 304.03, 304.73, 305.53, F11.11xx, or F11.21xx and procedure codes 94.45, 94.64–94.69, HZ2xxxx-3xxxx, HZ4xxxx, HZ5xxxx-6xxxx, HZ81xxx-82xxx, HZ84xxx-86xxx, HZ88xxx-89xxx, HZ91xxx-92xxx, HZ94xxx-96xxx, HZ98xxx-99xxx. This approach has been previously used by Peterson et al. [7]

Outcomes

Study outcomes were index hospitalization healthcare utilization and inpatient mortality. We assessed the length of hospital stay (above/below median), the total hospital charges in U.S. dollars (above/below median for each year) and the discharge disposition, i.e., to home vs. a rehabilitation facility, which included short- or long-term care hospital, skilled nursing facility, intermediate care facility, or a certified nursing facility. We also assessed the inpatient mortality during the index hospitalization.

We assessed several important covariate and potential confounders including socio-demographics (age, sex, race/ethnicity, income [in quartiles]), comorbidity (Deyo-Charlson comorbidity index, a validated measure that included 17 comorbidities, based on the presence of ICD-9-CM codes [12], categorized as 0, 1 and ≥2), insurance payer (Medicare, Medicaid, private, self-pay or other), and hospital characteristics. We categorized hospital location/teaching status as rural, urban non-teaching or urban teaching hospital; hospital bed size as small, medium or large, using the NIS cut-offs that vary by the year; and hospital region as Northeast, Midwest, South, and West.

Statistical analyses

We assessed summary statistics for the study cohort. We examined the time-trends by examining the rates of hospitalization for OUD as the primary diagnosis from 1998 to 2016 per 100,000 NIS claims. Inpatient mortality rates were similarly assessed over time for those hospitalized with OUD per 100,000 NIS claims and per 100,000 OUD claims.

We examined healthcare utilization outcomes over time. We performed separate multivariable-adjusted logistic regression analyses to assess the factors associated with each OUD hospitalization-related healthcare utilization outcome, i.e., the total hospital charges above/below the median, discharge to a rehabilitation facility vs. home, the length of hospital stay above/below the median and inpatient mortality. Models included all covariates and potential confounders of interest described in the section above. We calculated the odds ratios (OR) and 95% confidence intervals (CI).

Patient and public involvement

There was no direct patient involvement in the development of the study question or the execution of the study.

Results

Study cohort characteristics

For the study period from 1998 to 2016, we estimated a total of 781,767 OUD hospitalizations. The mean age was 43.7 years (standard error, 0.1), 52% were male, 67% White, and about a quarter each had Medicare, Medicaid or private insurance payer (Table 1). Fifty-four percent of people admitted for OUD were 45 years or younger. The majority (60%) were relatively healthy with a Deyo-Charlson score of zero. OUD hospitalizations were the highest in the lowest income classes; 30% in the first quartile and 27% in the second income quartile with the other quartiles near 20% each. We found that 2.4% of people hospitalized primarily for OUD died during hospitalization and 9.9% left against medical advice (Table 1).

Table 1

Characteristics of people with opioid use disorder (OUD)* hospitalizations in the U.S. from 1998–2016.

N (%), unless specified otherwise	Primary OUD-hospitalizationsN, projected** = 781,767
Age, Mean (SE); median	43.7 (0.10); 42.7
Age category, in years
<34	248,077 (31.74%)
34–45	177,702 (22.74%)
>45–55	167,100 (21.38%)
>55	188,645 (24.14%)
Sex
Male	407,778 (52.20%)
Female	373,377 (47.80%)
Race
White	520,536 (66.59%)
Black	65,868 (8.43%)
Hispanic	48,157 (6.16%)
Other/missing	147,149 (18.82%)
Deyo-Charlson Score
0	471,282 (60.28%)
1	164,829 (21.08%)
≥2	145,656 (18.63%)
Hospital Location/Teaching
Rural	88,052 (11.55%)
Urban nonteaching	315,210 (41.35%)
Urban teaching	359,059 (47.10%)
Insurance
Medicaid	222,798 (28.58%)
Medicare	212,004 (27.19%)
Other	42,709 (5.48%)
Private	179,884 (23.07%)
Self	122,215 (15.68%)
Income Category
First quartile	225,968 (29.83%)
Second quartile	202,685 (26.76%)
Third quartile	178,671 (23.59%)
Fourth quartile	150,090 (19.82%)
Hospital Bed size
Small	103,493 (13.58%)
Medium	210,779 (27.65%)
Large	448,050 (58.77%)
Hospital Region
Northeast	161,462 (21.10%)
Midwest	182,787 (23.88%)
South	277,347 (36.24%)
West	143,745 (18.78%)
Outcomes
Total Hospital Charge, Mean (SE); median, U.S. $	23,876 (314); 12,196
Discharge Status
Inpatient	161,826 (23.66%)
Home	522,243 (76.34%)
Length of Hospital Stay, Mean (SE); median	3.6 (0.02); 1.9
Length of Hospital Stay category, days***
≤3	571,442 (73.10%)
>3	210,325 (26.90%)
Died during hospitalization	18,394 (2.36%)
Discharge Against Medical Advice	77,323 (9.89%)

*Opioid drug abuse hospitalizations included those with primary diagnostic code of the following: 304.0x, 304.7x, 305.5x, 965.0x, E850.0, E935.0, F111.xxx, F112.xxx, T40.1X1x-4x,T40.2X1x-4x, or T40.3X1x-4x We excluded hospitalizations with ICD-9-CM diagnostic or procedure codes corresponding to drug/alcohol counseling and rehabilitation/detoxification including diagnostic codes 304.03, 304.73, 305.53, F11.11xx, or F11.21xx and procedure codes 94.45, 94.64–94.69, HZ2xxxx-3xxxx, HZ4xxxx, HZ5xxxx-6xxxx, HZ81xxx-82xxx, HZ84xxx-86xxx, HZ88xxx-89xxx, HZ91xxx-92xxx, HZ94xxx-96xxx, HZ98xxx-99xxx.

** Based on N, actual = 161,056

***The median hospital stay for all NIS hospitalizations was 3 days, which was used to categorize this variable

Characteristics of OUD-hospitalizations and outcomes by region

We found that compared to the Northeast, people with OUD-hospitalizations in the other 3 U.S. regions were more likely to be older, female, have Deyo-Charlson comorbidity index score ≥2, have Medicare, be admitted to a hospital with large bed size; and less likely to be White, have Medicaid, be in the highest income quartile, be admitted to urban, teaching hospital (Table 2).

Table 2

OUD-hospitalization characteristics by U.S. hospital region.

N (%), unless specified otherwise	NortheastN = 161,462 (21.10%)	MidwestN = 182,787 (23.88%)	South N = 277,347 (36.24%)	WestN = 143,745 (18.78%)
Age category, in years*
<34	57,795 (35.80%)	60,776 (33.25%)	88,103 (31.77%)	35,534(24.75%)
34–45	41,284 (25.75%)	45,645 (24.97%)	58,727 (21.18%)	28,298 (19.71%)
>45–55	32,326 (20.02%)	37,932 (20.75%)	60,897 (21.96%)	32,793 (22.84%)
>55	30,046 (18.61%)	38,424 (21.02%)	69,564 (25.09%)	46,953 (32.79%)
Sex*
Male	97,330 (60.30%)	95,865 (52.45%)	136,533 (49.24%)	69,581 (48.57%)
Female	64,087 (39.70%)	86,912 (47.55%)	140,751 (50.76%)	73,679 (51.43%)
Race*
White	113,563 (70.34%)	99,314 (54.33%)	204,933 (73.89%)	94,260 (65.59%)
Black	19,306 (11.96%)	16,676 (9.12%)	22,900 (8.26%)	6,471 (4.50%)
Hispanic	17,193 (10.65%)	2,290 (1.25%)	12,583 (4.54%)	15,444 (10.75%)
Other/missing	11,396 (7.06%)	64,502 (35.29%)	36,931 (13.32%)	27,545 (19.17%)
Deyo-Charlson Score*
0	103,460 (64.08%)	110,104 (60.24%)	165,652 (59.73%)	80,599 (56.07%)
1	32,067 (19.86%)	40,668 (22.25%)	58,226 (20.99%)	30,810 (21.43%)
≥2	25,935 (16.06%)	32,015 (17.51%)	53,469 (19.28%)	32,336 (22.50%)
Hospital Location/Teaching*
Rural	11,938 (7.39%)	21,216 (11.73%)	42,343 (15.30%)	12,555 (8.76%)
Urban nonteaching	46,363 (28.71%))	68,256 (37.73%)	122,488 (44.26%)	78,103 (54.52%)
Urban teaching	103,161 (63.89%)	91,414 (50.54%)	111,896 (40.44%)	52,590 (36.71%)
Insurance*
Medicaid	1,347,595 (52.11%)	668,596 (41.73%)	682,726 (30.95%)	455,310 (32.37%)
Medicare	389,372 (15.06%)	324,596 (20.26%)	510,635 (23.15%)	366,844 (26.00%)
Other	83,110 (3.21%)	88,816 (5.54%)	159,076 (7.21%)	117,542 (8.33%)
Private	430,734 (16.65%)	339,809 (21.21%)	422,857 (19.17%)	337,234 (23.90%)
Self	335,482 (12.97%)	180,332 (11.26%)	430,700 (19.52%)	134,087 (9.50%)
Income Category*
First quartile	34,756 (22.71%)	56,282 (31.30%)	99,603 (36.93%)	27,866 (20.20%)
Second quartile	32,055 (20.95%)	53,517 (29.57%)	78,213 (29.00%)	34,244 (24.82%)
Third quartile	38,096 (24.89%)	41,539 (22.95%)	56,275 (20.87%)	40,108 (29.07%)
Fourth quartile	48,125 (31.45%)	29,658 (16.39%)	35,592 (13.20%)	35,765 (25.92%)
Hospital Bed size*
Small	28,650 (17.74%)	24,854 (13.74%)	34,092 (12.32%)	15,897 (11.10%)
Medium	48,136 (29.81%)	43,233 (23.90%)	81,264 (29.37%)	38,146 (26.63%)
Large	84,676 (52.44%)	112,799 (62.36%)	161,370 (58.31%)	89,205 (62.27%)
Outcomes
Discharge Status*
Inpatient	31,564 (24.24%)	35,809 (22.27%)	61,265 (24.56%)	29,761 (22.99%)
Home	98,658 (75.76%)	125,022 (77.74%)	188,208 (75.44%)	99,695 (77.01%)
Length of Hospital Stay category, days*
≤3	103,047 (63.82%)	132,027 (72.73%)	183,251 (66.07%)	98,878 (68.79%)
>3	58,415 (36.18%)	50,760 (27.77%)	94,096 (33.93%)	44,867 (31.21%)
Length of Hospital Stay, Mean (SE)*; median	4.1 (0.07); 2.0	3.2 (0.05); 1.8	3.7 (0.03); 1.9	3.6 (0.04); 1.8
Total Hospital Charges, Mean (SE)*; median	24,463 (619); 11,418	17,828 (711); 9,714	23,134 (275); 12,361	34,403 (529); 18,852
Died during hospitalization*	4,083 (2.54%)	3,986 (2.18%)	6,368 (2.30%)	3,857 (2.69%)
Discharge Against Medical Advice*	26,534 (16.43%)	17,651 (9.66%)	21,246 (7.66%)	9,687 (6.74%)

Total N, projected for OUD-hospitalization = 781,767

*Statistically significantly with a p-value <0.001

Compared to the Northeast, we found that a slightly lower proportion of OUD hospitalizations in the other 3 U.S. regions had discharge to non-home settings, had hospital length of stay >3 days or left against medical advice (all with p-value <0.001; Table 2). Mean hospital stay was longest in the Northeast; mean hospital charges were the highest in the West followed by Northeast. Differences in in-hospital mortality were also statistically significant, but small in magnitude.

Outcomes of opioid use disorder hospitalizations by age, sex and race

We noted the people with older age >55 and females with OUD-hospitalization were significantly more likely than younger people and males to be discharged to non-home settings, have hospital charges higher than the median, or hospital stay >3 days (S1 Table). Whites were more likely to be discharged to non-home settings compared to all other race/ethnicities. Differences in mortality by age, sex and race were small.

Time-trends in opioid use disorder hospitalization and associated mortality and healthcare utilization

OUD hospitalizations were 59.8 per 100,000 of all NIS hospitalizations in the U.S. with any diagnosis in 1998–2000, which increased steadily over the study period to 190.7 per 100,000 NIS hospitalizations in 2015–16, i.e. a 219% increase, leading to a rate increase of 3.2-fold (Table 3; Fig 1). The mortality rate for OUD hospitalization was 1.2 per 100,000 NIS hospitalizations in the U.S. in 1998–2000 that increased 5-times to 5.9 per 100,000 NIS hospitalizations in 2015–16 (Table 3).

Table 3

Time-trends in OUD hospitalization and mortality rates from 1998 to 2016 and the comparative non-OUD mortality rates.

	Total NIS claims	OUD claims	OUD deaths	OUD claims Per 100K total NIS claims	OUD Death rate Per 100K NIS claims	OUD Death rate per 1k primary OUD claims	Comparative Death rate per 1k Non-OUD claims
1998–2000	103,665,051	62,010	1,226	59.82	1.18	19.77	23.79
2001–2002	72,617,381	53,176	1,002	73.23	1.38	18.84	22.18
2003–2004	74,571,583	63,853	1,228	85.63	1.65	19.23	20.88
2005–2006	75,919,595	66,923	1,350	88.15	1.78	20.17	19.73
2007–2008	76,366,797	78,541	1,524	102.85	2.00	19.40	19.07
2009–2010	75,086,597	97,611	2,149	130.00	2.86	22.02	18.24
2011–2012	73,447,261	112,428	2,445	153.07	3.33	21.75	18.07
2013–2014	70,956,610	110,985	3,255	156.41	4.59	29.33	18.93
2015–2016	71,445,363	136,240	4,215	190.69	5.90	30.94	19.05

All rates are expressed per 100k or per 1k claims or hospitalizations

The last column represents the death rate in all NIS hospitalizations except OUD hospitalizations.

Fig 1

Time-trend in OUD hospitalization rate per 100,000 NIS claims from 1998 to 2016.

X-axis represents time-periods from 1998–2000 to 2015–16. Y-axis shows primary OUD hospitalization rates per 100,00 NIS claims.

Among the OUD hospitalizations, the mortality rate increased from 19.8 per 1,000 OUD hospitalizations in 1998–2000 to 30.9 per 1,000 OUD hospitalizations in 2015–16 (Table 3). In comparison, mortality rate decreased for non-OUD hospitalizations over the same period from 23.8 to 19 per 1,000 non-OUD hospitalizations (Fig 2). Time-related increase in OUD hospitalizations and associated mortality was seen in all age groups, both sexes and in both white and non-white race/ethnicity (data available on request).

Fig 2

Comparison of OUD vs. non-OUD death rates over the study period from 1998 to 2016.

X-axis represents time-periods from 1998–2000 to 2015–16. Y-axis shows the in-hospital death rates per 1k among primary OUD hospitalizations (hashed bars) and among all NIS hospitalizations except OUD (solid bars).

Time-trends in OUD hospitalization associated healthcare utilization showed an increase over the study period, with the mean (median) total hospital charges increased from $8,261 ($4,339) to $32,792 ($18,244; Table 3). In contrast, we saw little change in mean (median) length of hospital stay from 3.2 days (1.6) to 3.9 days (2.2) over the study period and no change in the proportion discharged to home, i.e., 80% in 1998–2000 versus 80% in 2015–16 (Table 4).

Table 4

Time-trends in healthcare utilization outcomes for OUD Hospitalizations from 1998 to 2016.

	Total hospital charges, US $	Discharged home	Length of Hospital Stay, days
	Mean (SE); median	N (%)	Mean (SE); median
1998–2000	8,261 (601); 4,339	39,726 (80%)	3.2 (0.13); 1.6
2001–2002	11,101 (729); 5,676	33,074 (75%)	3.4 (0.07); 1.8
2003–2004	13,830 (1,003); 7,064	41,975 (76%)	3.4 (0.12); 1.7
2005–2006	17,756 (557); 9,573	44,941 (75%)	3.6 (0.07); 1.8
2007–2008	22,767 (563); 11,912	52,091 (74%)	3.7 (0.08); 1.8
2009–2010	24,844 (648); 13,387	65,752 (75%)	3.7 (0.06); 1.9
2011–2012	28,210 (573); 15,634	75,989 (75%)	3.7 (0.05); 1.9
2013–2014	32,666 (477); 18,188	73,985 (75%)	3.8 (0.04); 1.9
2015–2016	32,792 (586); 18,244	94,710 (80%)	3.9 (0.05); 2.2

SE, standard error; US $, US dollar

Multivariable-adjusted predictors of healthcare utilization and inpatient mortality in people admitted with opioid use disorder

In the multivariable-adjusted analysis, compared to age <34 years, older age was associated with a higher risk of hospital charges above the median and the length of hospital stay >3 days and a slightly higher risk of discharge to a rehabilitation facility (Table 5). Higher Deyo-Charlson score was associated with higher hospital charges, a longer length of hospital stay, and higher inpatient mortality (Table 5). Women had 26% higher odds and Blacks 31% lower odds of discharge to a rehabilitation facility, compared to men and Whites, respectively.

Table 5

Predictors of healthcare utilization for people with an OUD hospitalization in the U.S.

	Hospital charge above the median*	Length of hospital stay > 3 days	Discharge to inpatient facility	In-hospital Mortality
Age category
<34 years	Ref	Ref	Ref	Ref
34–45 years	1.24 (1.20, 1.28)	1.06 (1.02, 1.09)	1.05 (1.01, 1.09)	0.95 (0.86, 1.04)
>45–55 years	1.61 (1.55, 1.66)	1.21 (1.17, 1.26)	1.02 (0.98, 1.06)	0.82 (0.74, 0.91)
>55 years	1.72 (1.66, 1.79)	1.41 (1.36, 1.47)	1.16 (1.11, 1.21)	0.74 (0.66, 0.83)
Sex
Male	Ref	Ref	Ref	Ref
Female	0.97 (0.95, 0.99)	1.01 (0.99, 1.04)	1.26 (1.23, 1.30)	0.75 (0.70, 0.81)
Race
White	Ref	Ref	Ref	Ref
Black	0.87 (0.84, 0.91)	0.84 (0.81, 0.88)	0.69 (0.65, 0.73)	0.67 (0.59, 0.77)
Hispanic	1.13 (1.07, 1.18)	0.91 (0.86, 0.95)	0.83 (0.78, 0.88)	0.76 (0.66, 0.88)
Other/missing	0.84 (0.82, 0.87)	0.77 (0.74, 0.79)	0.88 (0.85, 0.91)	0.90 (0.82, 0.99)
Deyo-Charlson comorbidity Score
0	Ref	Ref	Ref	Ref
1	1.57 (1.53, 1.62)	1.35 (1.31, 1.39)	0.92 (0.89, 0.95)	1.84 (1.69, 2.02)
≥2	2.16 (2.09, 2.23)	1.91 (1.84, 1.97)	1.05 (1.01, 1.09)	2.46 (2.23, 2.72)
Insurance
Private	Ref	Ref	Ref	Ref
Medicaid	0.97 (0.94, 1.00)	1.06 (1.03, 1.10)	0.82 (0.79, 0.85)	1.17 (1.06, 1.29)
Medicare	1.15 (1.11, 1.19)	1.11 (1.07, 1.15)	1.20 (1.16, 1.25)	0.79 (0.71, 0.88)
Other	1.08 (1.03, 1.14)	1.06 (1.00, 1.12)	0.89 (0.84, 0.95)	0.98 (0.83, 1.15)
Self	1.04 (1.01, 1.08)	0.72 (0.69, 0.75)	0.75 (0.72, 0.79)	1.24 (1.11, 1.38)
Income category
First quartile	0.86 (0.83, 0.89)	1.06 (1.02, 1.10)	0.79 (0.76, 0.83)	0.96 (0.87, 1.06)
Second quartile	0.86 (0.83, 0.89)	1.00 (0.96, 1.03)	0.88 (0.85, 0.92)	0.97 (0.88, 1.07)
Third quartile	0.93 (0.90, 0.96)	1.02 (0.98, 1.06)	0.92 (0.89, 0.96)	0.98 (0.89, 1.08)
Fourth quartile	Ref	Ref	Ref	Ref
Hospital region
Northeast	Ref	Ref	Ref	Ref
Midwest	0.58 (0.56, 0.60)	0.71 (0.69, 0.74)	0.88 (0.85, 0.92)	0.91 (0.82, 1.01)
South	0.90 (0.88, 0.93)	0.93 (0.90, 0.97)	0.95 (0.91, 0.98)	1.00 (0.91, 1.11)
West	1.66 (1.60, 1.72)	0.75 (0.72, 0.77)	0.82 (0.78, 0.85)	1.16 (1.04, 1.29)
Hospital teaching status
Rural	Ref	Ref	Ref	Ref
Urban nonteaching	2.28 (2.19, 2.37)	1.42 (1.36, 1.48)	0.98 (0.94, 1.03)	1.57 (1.36, 1.82)
Urban teaching	2.42 (2.32, 2.52)	1.81 (1.74, 1.89)	0.84 (0.80, 0.87)	2.17 (1.88, 2.50)
Hospital bed size
Small	Ref	Ref	Ref	Ref
Medium	1.35 (1.30, 1.40)	1.08 (1.04, 1.12)	1.02 (0.98, 1.07)	1.25 (1.11, 1.40)
Large	1.53 (1.48, 1.58)	1.24 (1.20, 1.29)	1.00 (0.96, 1.04)	1.19 (1.07, 1.32)

*Total hospital charge were categorized as above or below the median for each year individually

We found that women had 25% lower odds of inpatient mortality than men, blacks had 33% lower odds of mortality than whites and older age was associated with higher inpatient mortality. The models for hospital charges, length of stay, and inpatient mortality also showed better outcomes for rural hospitals compared with both urban teaching and urban non-teaching hospitals. Compared to the hospitals in the Northeast U.S., those in the Midwest and the South had lower hospital charges, shorter length of stay, and lower odds of discharge to non-home settings. Lower income was associated with lower hospital charges and lower odds of discharge to non-home settings (Table 5).

Discussion

We performed a longitudinal study of OUD hospitalizations over a 19-year period from 1998 to 2016, the most recent year of publicly available NIS data. We examined the time-trends in OUD hospitalizations and associated healthcare utilization outcomes and mortality, and their predictors. Our multivariable-adjusted models identified several factors independently associated with each healthcare utilization and in-hospital mortality, while all the other factors shown were adjusted for in the analyses. Several findings of this study merit further discussion.

The OUD hospitalizations in the U.S. increased steadily from 62,010 in 1998–2000 to 136,240 in 2015–2016, the most recent period with available data. We noted a 219% increase in OUD hospitalizations, compared to the baseline from 1998–2000. The continued rise in OUD hospitalizations in the U.S. is of concern. State and federal agencies have implemented several policies for OUD and various programs to reduce related morbidity [10,13-16]. This increasing trend in OUD hospitalizations in the U.S. confirms the impact of OUD epidemic on the healthcare system, and describes the magnitude of the problem. These findings are also consistent with an increasing OUD in delivery hospitalizations to 2014 [17].

The death rate for OUD hospitalizations was 1.2 per 100,000 NIS hospitalizations in 1998–2000 that increased 5-times to 5.9 per 100,000 NIS hospitalizations in 2015–16. The increase in the OUD hospitalization mortality rate continued through the most recent study period, 2015–16. There was an increase of 77% between 2011–2012 to 2015–16. This is consistent with national CDC estimates of rapidly increasing OUD-related deaths, noted to be 28,647, 33,091 and 42,249 in 2014, 2015 and 2016, respectively [1,6]. The 47% increase in OUD-related mortality from 2014 to 2016 was alarming [6].

Compared to mortality rate in the general population with hospitalization, OUD-related mortality rates were 0.8 times in 1998–2000, but rose to 1.6 times higher in 2015–2016. This indicates worsening of the mortality outcome in OUD-related hospitalizations over time, relative to all other hospitalizations in the U.S. This might be related to a higher severity of opioid abuse, a reduction in access to care or higher associated psychiatric or medical comorbidity over time. These hypotheses need further examination.

In unadjusted comparisons, we noted the OUD hospitalizations in people with older age >55, females and Northeast U.S. region had higher healthcare utilization; OUD hospitalizations in Northeast were also associated with higher proportion of people leaving against medical advice. Whites had higher rate of discharge to non-home settings after OUD hospitalizations compared to other race/ethnicities. Mortality rates were only slightly different by any of these characteristics. We also found interesting differences in patient characteristics by U.S. region in OUD hospitalizations.

The implementation of effective policy and public health programs in the U.S. has the potential to reverse the trend in OUD hospitalizations in the near future [9,10,13-16,18-20]. Strategies and programs to reduce OUD and improve OUD outcomes are being developed. Examples include a system-wide organizational opioid stewardship program (OSP) that was associated with a reduction in opioid morbidity [21]. A combined implementation of mandated provider review of state-run prescription drug monitoring program and pain clinic laws reduced opioid amounts prescribed by 8% and prescription opioid overdose death rates by 12% [22]. Telemedicine has the potential to improve the provision of evidence-based medication-assisted treatment for OUD [23]. The use of buprenorphine and methadone maintenance treatment after non-fatal opioid overdose reduced all-cause and opioid-related mortality [24]. This finding is supported by a systematic review and meta-analysis of 19 studies with 138,716 people treated with either methadone or buprenorphine for opioid dependence [25]. Thus, effective strategies exist to reduce the OUD-related morbidity and mortality.

We examined important patient/clinical characteristics associated with OUD hospitalization related healthcare utilization and mortality. Older age, White race, a higher Deyo-Charlson score and female sex were each associated with worse healthcare utilization outcomes or mortality related to index OUD hospitalization. A previous CDC analysis of drug overdose deaths (prescription opioids and heroin were the main causes) using the 2013 and 2014 national data found that age-adjusted mortality rates for Whites, Blacks and Hispanics were 19, 10.5 and 6.7 per 100,000 [3]. In a study of OUD-hospitalization mortality, Whites, ages 50–64, Medicare beneficiaries with disabilities, and residents of lower-income areas were noted to have higher odds of opioid/heroin poisoning [26]. These studies provide one potential reason for higher mortality in Whites and are consistent with our observation of an independent association of White race with higher mortality during OUD-hospitalization, adjusted for age, sex, insurance, income, comorbidity, hospital region (rural/urban) and teaching status, location or bed size. Future studies are needed to assess the other underlying causes for higher mortality in Whites with OUD-hospitalizations.

Our observation of the association of male sex with higher inpatient mortality of OUD-hospitalization extends similar observations in people who underwent elective total joint replacement [27] or with pharmaceutical opioid related overdose deaths [28]. We also noted differences by region and by hospital characteristics in these outcomes, which extend similar findings for opioid and heroin-related overdose hospitalizations [8] to OUD-related hospitalizations.

Our study findings must be interpreted considering study limitations. Misclassification bias is possible, since we used diagnostic codes for the identification of the study cohort and comorbidities. Our observational cohort study design puts this study at the potential risk of residual confounding for the predictors of healthcare utilization and mortality outcomes; we adjusted for multiple covariates and confounders to reduce the risk of confounding bias. We assessed hospital charges, which are usually inflated and do not reflect the actual cost of the hospitalization. Due to the lack of cause of death data in the NIS, we are unable to comment on whether the causes of death changed over time, were attributed to OUD or related disorder (hepatitis C, HIV, endocarditis, valvular disease) or differed by factors significantly associated with higher mortality. Longer-term studies of mortality up to 4 years after OUD hospitalization found that both opioid use and physical comorbidities contributed to mortality [29,30].

Our study has many strengths. We used the U.S. NIS, a national dataset that makes our results generalizable to the general U.S. population. We used two decades of data to examine the time-trends in OUD hospitalization, another study strength.

Conclusions

In conclusion, we found increasing rates of OUD hospitalizations and OUD mortality rates from 1998 to 2016. These time-trends are concerning, given the alarmingly high rates of associated mortality and no trends of a slow-down or decline. We identified factors associated with healthcare utilization and mortality outcomes for OUD hospitalizations. Future studies need to examine the most effective strategies to reduce OUD hospitalizations and associated mortality and healthcare utilization.

OUD-hospitalization outcomes by age, sex and race.

(DOCX)

Click here for additional data file.

Transfer Alert

This paper was transferred from another journal. As a result, its full editorial history (including decision letters, peer reviews and author responses) may not be present.

20 Nov 2019

PONE-D-19-24166

National U.S. Time-trends in Opioid Use Disorder Hospitalizations and Associated Healthcare Utilization and Mortality

PLOS ONE

Dear Dr Singh,

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.

We would appreciate receiving your revised manuscript by Jan 04 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,

Sandra C. Buttigieg, MD PhD FFPH MSc MBA MMCFD

Academic Editor

PLOS ONE

Journal Requirements:

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

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.journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and http://www.journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

2. We note that you have indicated that data from this study are available upon request. PLOS only allows data to be available upon request if there are legal or ethical restrictions on sharing data publicly. For information on unacceptable data access restrictions, please see http://journals.plos.org/plosone/s/data-availability#loc-unacceptable-data-access-restrictions.

* In your revised cover letter, please address the following prompts:

a) If there are ethical or legal restrictions on sharing a de-identified data set, please explain them in detail (e.g., data contain potentially identifying or sensitive patient information) and who has imposed them (e.g., an ethics committee). Please also provide contact information for a data access committee, ethics committee, or other institutional body to which data requests may be sent.

b) If there are no restrictions, please upload the minimal anonymized data set necessary to replicate your study findings as either Supporting Information files or to a stable, public repository and provide us with the relevant URLs, DOIs, or accession numbers. Please see http://www.bmj.com/content/340/bmj.c181.long for guidelines on how to de-identify and prepare clinical data for publication. For a list of acceptable repositories, please see http://journals.plos.org/plosone/s/data-availability#loc-recommended-repositories.

We will update your Data Availability statement on your behalf to reflect the information you provide.

3. Thank you for stating the following in the Competing Interests section:

"I have read the journal's policy and the authors of this manuscript have the following competing interests. JAS has received consultant fees from Crealta/Horizon, Medisys, Fidia, UBM LLC, Medscape, WebMD, Clinical Care options, Clearview healthcare partners, Putnam associates, Spherix, the National Institutes of Health and the American College of Rheumatology.  JAS owns stock options in Amarin pharmaceuticals and Viking therapeutics. JAS is a member of the executive of OMERACT, an organization that develops outcome measures in rheumatology and receives arms-length funding from 36 companies. JAS serves on the FDA Arthritis Advisory Committee. JAS is a member of the Veterans Affairs Rheumatology Field Advisory Committee. JAS is the editor and the Director of the UAB Cochrane Musculoskeletal Group Satellite Center on Network Meta-analysis.  JAS previously served as a member of the following committees: member, the American College of Rheumatology's (ACR) Annual Meeting Planning Committee (AMPC) and Quality of Care Committees, the Chair of the ACR Meet-the-Professor, Workshop and Study Group Subcommittee and the co-chair of the ACR Criteria and Response Criteria subcommittee.  JDC has no conflicts. ".

i) Please confirm that this does not alter your adherence to all PLOS ONE policies on sharing data and materials, by including the following statement: "This does not alter our adherence to  PLOS ONE policies on sharing data and materials.” (as detailed online in our guide for authors http://journals.plos.org/plosone/s/competing-interests).  If there are restrictions on sharing of data and/or materials, please state these. Please note that we cannot proceed with consideration of your article until this information has been declared.

ii) Please include your updated Competing Interests statement in your cover letter; we will change the online submission form on your behalf.

Please know it is PLOS ONE policy for corresponding authors to declare, on behalf of all authors, all potential competing interests for the purposes of transparency. PLOS defines a competing interest as anything that interferes with, or could reasonably be perceived as interfering with, the full and objective presentation, peer review, editorial decision-making, or publication of research or non-research articles submitted to one of the journals. Competing interests can be financial or non-financial, professional, or personal. Competing interests can arise in relationship to an organization or another person. Please follow this link to our website for more details on competing interests: http://journals.plos.org/plosone/s/competing-interests

Additional Editor Comments (if provided):

I have read this manuscript with great interest. I advise you to take note of the very detailed comments of both reviewers and respond accordingly. I look forward to your revised manuscript.

[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: Yes

Reviewer #2: Yes

**********

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

Reviewer #1: Yes

Reviewer #2: Yes

**********

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: No

Reviewer #2: Yes

**********

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: Yes

Reviewer #2: Yes

**********

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: The authors used 781,000+ hospital records from a National Institutional sample of hospitalizations for the period 1998 through 2016 to calculate rates of hospitalization by age race and gender for Opioid overdose Hospitalizations and in hospital mortality noting that the rates increase from 59.8 per 100,00 to 190 per 100,000 over the 18 years period.

They indicate that on a national basis this is one of the first and only papers to chronicle this information which makes it unique. They examined healthcare utilization outcomes over time. Can they move this forward by examining some of the regional differences in west, midwest, Northeast and south regions , are these all the same with regard to white and black and by age?

A geospatial map of rates or table would be helpful showing black , white Hospitalization rates as well as a regional difference map as opposed to a multivariate logistic regression which is interesting but does not show important nuances in time , person and place.

Much can be learned in the details of this study, Older white male rates of mortality and hospitalization have been detailed previously so presenting some of the details would prove to be important.

A few questions, they indicate older white opioid hospitalizations fared worse that black and younger admits. Please comment on the reasons for this , specifically what were the causes of death among older white admits ; was this just an age phenomenon or was there something specific to region or urban /rural.

Please show stratification by average days of hospitalization by age and gender and race.

OUD and old age can be linked to end stage cancer diagnoses, please provide the information of end of life related causes of death versus others.

How much of the co morbidity was opioid and drug related, aka, Hepatitis C, HIV and myocardial endocarditis and valvular disease?

ALso the logistic regressions should be given more detail with tables shown. What was the median length of stay hospitalization for those under 55 and over 55?

This data set has the ability to show important ground breaking information .

Reviewer #2: The authors conducted an observational study form National Inpatient Sample, and showed an increase in OUD hospitalizations and associated inpatient mortality. Age, sex, race, and Deyo-Charlson comorbidity index are independent factors of healthcare utilization and mortality related to index OUD hospitalization.

Is there any data to support these independent factors associated with healthcare utilization and mortality related to OUD hospitalization?

Some comorbidity indices have been developed to measure and weigh the overall burden of comorbidities, for example, Charlson comorbidity index(CCI), Deyo CCI, Romano CCI. Is there any special reason for choosing Deyo CCI in this article.

Minor point: Table 4: The form after “Rural” is empty. Should it be “Ref”?

**********

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

[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.

27 Nov 2019

We thank the reviewers for their comments. Following are our point-by-point responses to their comments.

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: Yes

Reviewer #2: Yes

Response: Thank you.

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

Reviewer #1: Yes

Reviewer #2: Yes

Response: Thank you.

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: No

Reviewer #2: Yes

Response: Thank you. AHRQ data have to be obtained directly from AHRQ. AHRQ does not allow individual researcher to disseminate the data. These de-identified data are publically available. We had included this statemnet in the data sharing section.

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: Yes

Reviewer #2: Yes

Response: Thank you

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: The authors used 781,000+ hospital records from a National Institutional sample of hospitalizations for the period 1998 through 2016 to calculate rates of hospitalization by age race and gender for Opioid overdose Hospitalizations and in hospital mortality noting that the rates increase from 59.8 per 100,00 to 190 per 100,000 over the 18 years period.

They indicate that on a national basis this is one of the first and only papers to chronicle this information which makes it unique. They examined healthcare utilization outcomes over time. Can they move this forward by examining some of the regional differences in west, midwest, Northeast and south regions , are these all the same with regard to white and black and by age?

Response: We have provided this as a new table as suggested, and have these results to results and discussion sections.

Table 2. OUD-hospitalization characteristics by U.S. hospital region

N (%), unless specified otherwise Northeast

N=161,462 (21.10%)

Midwest

N=182,787 (23.88%) South N=277,347 (36.24%) West

N=143,745 (18.78%)

Age category, in years*

<34 57,795 (35.80%) 60,776 (33.25%) 88,103 (31.77%) 35,534(24.75%)

34 - 45 41,284 (25.75%) 45,645 (24.97%) 58,727 (21.18%) 28,298 (19.71%)

>45 - 55 32,326 (20.02%) 37,932 (20.75%) 60,897 (21.96%) 32,793 (22.84%)

>55 30,046 (18.61%) 38,424 (21.02%) 69,564 (25.09%) 46,953 (32.79%)

Sex*

Male 97,330 (60.30%) 95,865 (52.45%) 136,533 (49.24%) 69,581 (48.57%)

Female 64,087 (39.70%) 86,912 (47.55%) 140,751 (50.76%) 73,679 (51.43%)

Race*

White 113,563 (70.34%) 99,314 (54.33%) 204,933 (73.89%) 94,260 (65.59%)

Black 19,306 (11.96%) 16,676 (9.12%) 22,900 (8.26%) 6,471 (4.50%)

Hispanic 17,193 (10.65%) 2,290 (1.25%) 12,583 (4.54%) 15,444 (10.75%)

Other/missing 11,396 (7.06%) 64,502 (35.29%) 36,931 (13.32%) 27,545 (19.17%)

Deyo-Charlson Score*

0 103,460 (64.08%) 110,104 (60.24%) 165,652 (59.73%) 80,599 (56.07%)

1 32,067 (19.86%) 40,668 (22.25%) 58,226 (20.99%) 30,810 (21.43%)

≥2 25,935 (16.06%) 32,015 (17.51%) 53,469 (19.28%) 32,336 (22.50%)

Hospital Location/Teaching*

Rural 11,938 (7.39%) 21,216 (11.73%) 42,343 (15.30%) 12,555 (8.76%)

Urban nonteaching 46,363 (28.71%)) 68,256 (37.73%) 122,488 (44.26%) 78,103 (54.52%)

Urban teaching 103,161 (63.89%) 91,414 (50.54%) 111,896 (40.44%) 52,590 (36.71%)

Insurance*

Medicaid 1,347,595 (52.11%) 668,596 (41.73%) 682,726 (30.95%) 455,310 (32.37%)

Medicare 389,372 (15.06%) 324,596 (20.26%) 510,635 (23.15%) 366,844 (26.00%)

Other 83,110 (3.21%) 88,816 (5.54%) 159,076 (7.21%) 117,542 (8.33%)

Private 430,734 (16.65%) 339,809 (21.21%) 422,857 (19.17%) 337,234 (23.90%)

Self 335,482 (12.97%) 180,332 (11.26%) 430,700 (19.52%) 134,087 (9.50%)

Income Category*

First quartile 34,756 (22.71%) 56,282 (31.30%) 99,603 (36.93%) 27,866 (20.20%)

Second quartile 32,055 (20.95%) 53,517 (29.57%) 78,213 (29.00%) 34,244 (24.82%)

Third quartile 38,096 (24.89%) 41,539 (22.95%) 56,275 (20.87%) 40,108 (29.07%)

Fourth quartile 48,125 (31.45%) 29,658 (16.39%) 35,592 (13.20%) 35,765 (25.92%)

Hospital Bed size*

Small 28,650 (17.74%) 24,854 (13.74%) 34,092 (12.32%) 15,897 (11.10%)

Medium 48,136 (29.81%) 43,233 (23.90%) 81,264 (29.37%) 38,146 (26.63%)

Large 84,676 (52.44%) 112,799 (62.36%) 161,370 (58.31%) 89,205 (62.27%)

Outcomes

Discharge Status*

Inpatient 31,564 (24.24%) 35,809 (22.27%) 61,265 (24.56%) 29,761 (22.99%)

Home 98,658 (75.76%) 125,022 (77.74%) 188,208 (75.44%) 99,695 (77.01%)

Length of Hospital Stay category, days*

≤3 103,047 (63.82%) 132,027 (72.73%) 183,251 (66.07%) 98,878 (68.79%)

>3 58,415 (36.18%) 50,760 (27.77%) 94,096 (33.93%) 44,867 (31.21%)

Died during hospitalization* 4,083 (2.54%) 3,986 (2.18%) 6,368 (2.30%) 3,857 (2.69%)

Discharge Against Medical Advice* 26,534 (16.43%) 17,651 (9.66%) 21,246 (7.66%) 9,687 (6.74%)

Results

“Characteristics of OUD-hospitalizations and outcomes by Region

We found that compared to the Northeast, people with OUD-hospitalizations in the other 3 U.S. regions were more likely to be older, female, have Deyo-Charlson comorbidity index score ≥2, have Medicare, be admitted to a hospital with large bed size; and less likely to be White, have Medicaid, be in the highest income quartile, be admitted to urban, teaching hospital (Table 2).

Compared to the Northeast, we found that a slightly lower proportion of OUD hospitalizations in the other 3 U.S. regions had discharge to non-home settings, had hospital length of stay >3 days or left against medical advice (all with p-value <0.001; Table 2). Differences in in-hospital mortality were also statistically significant, but small in magnitude.”

Discussion

“In unadjusted comparisons, we noted the OUD hospitalizations in people with older age >55, females and Northeast U.S. region had higher healthcare utilization; OUD hospitalizations in Northeast were also associated with higher proportion of people leaving against medical advice. Whites had higher rate of discharge to non-home settings after OUD hospitalizations compared to other race/ethnicities. Mortality rates were only slightly different by any of these characteristics. We also found interesting differences in patient characteristics by U.S. region in OUD hospitalizations.”

A geospatial map of rates or table would be helpful showing black , white Hospitalization rates as well as a regional difference map as opposed to a multivariate logistic regression which is interesting but does not show important nuances in time , person and place.

Much can be learned in the details of this study, Older white male rates of mortality and hospitalization have been detailed previously so presenting some of the details would prove to be important.

Response: We have provided these data as requested as a new appendix with regards to age, race and sex with time-trends. We have added text to results and discussion.

“Outcomes of OUD hospitalizations by age, sex and race

We noted the people with older age >55 and females with OUD-hospitalization were significantly more likely than younger people and males to be discharged to non-home settings, have hospital charges higher than the median, or hospital stay >3 days (Appendix 1). Whites were more likely to be discharged to non-home settings compared to all other race/ethnicities. Differences in mortality by age, sex and race were small.

Appendix 1. OUD-hospitalization outcomes by age, sex and race

N (%), unless specified otherwise % Discharged to non-home settings Total hospital charges

>median Length of hospital stay >3 days Died during hospitalization

Age category, in years

<34 43,787 (21.05%) 73,411 (29.59%) 64,217 (25.89%) 5,963 (2.41%)

34 - 45 33,663 (22.38%) 63,307 (35.63%) 51,860 (29.18%) 4,284 (2.42%)

>45 - 55 34.309 (23.05%) 76,215 (45.61%) 58,458 (34.98%) 3,903 (2.34%)

>55 50,051 (28.33%) 99,568 (52.78%) 78,813 (41.78%) 4,229 (2.24%)

Sex

Male 73,108 (21.10%) 159,035 (39.00%) 127,738 (31.33%) 10,871 (2.67%)

Female 88,638 (26.29%) 153,372 (41.08%) 125,518 (33.62%) 7,523 (2.02%)

Race

White 116,236 (25.32%) 215,485 (41.40%) 173,431 (33.32%) 12,770 (2.46%)

Black 10,004 (17.57%) 26,722 (40.57%) 22,691 (34.45%) 1,349 (2.05%)

Hispanic 7,759 (19.12%) 22,602 (46.93%) 15,801 (32.82%) 1,154 (2.40%)

Other/missing 27,822 (17.19%) 47,759 (32.46%) 41,489 (28.20%) 3,116 (2.12%)

A few questions, they indicate older white opioid hospitalizations fared worse that black and younger admits. Please comment on the reasons for this , specifically what were the causes of death among older white admits ; was this just an age phenomenon or was there something specific to region or urban /rural.

Response: We have more discussion related to this association noted between race and in-hospital mortality. However, due to non-availability of cause of death in this dataset, we are unable to provide further insights into causes of death. Since these analyses were adjusted for age, sex, insurance, income, comorbidity, hospital region (rural/urban) and teaching status, location or bed size, none of the noted difference by race can be attributed to these factors.

“A previous CDC analysis of drug overdose deaths (prescription opioids and heroin were the main causes) using the 2013 and 2014 national data found that age-adjusted mortality rates for Whites, Blacks and Hispanics were 19, 10.5 and 6.7 per 100,000 [3]. This previous observation is consistent with our observation of an independent association of White race with higher mortality during OUD-hospitalization, adjusted for age, sex, insurance, income, comorbidity, hospital region (rural/urban) and teaching status, location or bed size. Future studies are needed to assess the underlying causes for higher mortality for Whites in OUD-hospitalizations.”

Please show stratification by average days of hospitalization by age and gender and race.

Response: Please see the new table and response added in response to an earlier comment by the reviewer. The length of stay >3 days seemed to more in older people, therefore the association of age with LOS was noted in OUD-hospitalization. We noted smaller differences in the length of hospitalization by sex and race.

OUD and old age can be linked to end stage cancer diagnoses, please provide the information of end of life related causes of death versus others.

Response: We do not have cause of death data in the NIS, therefore we can not definitively assess the cause of OUD-in-hospital death. As can be seen from the table above added per the reviewer request, most of the in-hospital mortality were in the younger people, i.e., 45% of deaths were in people <45 years and 77% of the deaths were in people <55 years. Therefore, it is unlikely that cancer is one of the main reasons for OUD-hospitalization associated death. It is more likely that OUD mortality is related to OUD- and associated complications, including medical complications.

How much of the co morbidity was opioid and drug related, aka, Hepatitis C, HIV and myocardial endocarditis and valvular disease?

Response: The NIS does not provide the cause of death, so we are unable to ascertain the attribution of death to medical comorbidities listed in the comment vs. the OUD itself. We have added this to the study limitations section.

“Due to the lack of cause of death data in the NIS, we are unable to comment on whether the causes of death changed over time, were attributed to OUD or related disorder (hepatitis C, HIV, endocarditis, valvular disease) or differed by factors significantly associated with higher mortality.”

ALso the logistic regressions should be given more detail with tables shown. What was the median length of stay hospitalization for those under 55 and over 55?

This data set has the ability to show important ground breaking information .

Response: We have provided additional data as suggested in response to an earlier comment to show data by age, sex and race. This has been added as an appendix/table. In addition, we have added more detail to the results section related to the logistic regression results.

“In the multivariable-adjusted analysis, compared to age <34 years, older age was associated with a higher risk of hospital charges above the median and the length of hospital stay >3 days and a slightly higher risk of discharge to a rehabilitation facility (Table 4). Higher Deyo-Charlson score was associated with higher hospital charges, a longer length of hospital stay, and higher inpatient mortality (Table 4). Women had 26% higher odds and Blacks 31% lower odds of discharge to a rehabilitation facility, compared to men and Whites, respectively.

We found that women had 25% lower odds of inpatient mortality than men, blacks had 33% lower odds of mortality than whites and older age was associated with higher inpatient mortality. The models for hospital charges, length of stay, and inpatient mortality also showed better outcomes for rural hospitals compared with both urban teaching and urban non-teaching hospitals . Compared to the hospitals in the Northeast U.S., those in the Midwest and the South had lower hospital charges, shorter length of stay, and lower odds of discharge to non-home settings. Lower income was associated with lower hospital charges and lower odds of discharge to non-home settings (Table 4).”

Discussion

“We performed a longitudinal study of OUD hospitalizations over a 19-year period from 1998 to 2016, the most recent year of publicly available NIS data. We examined the time-trends in OUD hospitalizations and associated healthcare utilization outcomes and mortality, and their predictors. Our multivariable-adjusted models identified several factors independently associated with each healthcare utilization and in-hospital mortality, while all the other factors shown were adjusted for in the analyses.”

Reviewer #2: The authors conducted an observational study form National Inpatient Sample, and showed an increase in OUD hospitalizations and associated inpatient mortality. Age, sex, race, and Deyo-Charlson comorbidity index are independent factors of healthcare utilization and mortality related to index OUD hospitalization.

Is there any data to support these independent factors associated with healthcare utilization and mortality related to OUD hospitalization?

Response: We have expanded the discussion related to associations we noted, that support our findings. In general, many findings were from smaller sub-cohorts of people with OUD.

“We examined important patient/clinical characteristics associated with OUD hospitalization related healthcare utilization and mortality. Older age, White race, a higher Deyo-Charlson score and female sex were each associated with worse healthcare utilization outcomes or mortality related to index OUD hospitalization. A previous CDC analysis of drug overdose deaths (prescription opioids and heroin were the main causes) using the 2013 and 2014 national data found that age-adjusted mortality rates for Whites, Blacks and Hispanics were 19, 10.5 and 6.7 per 100,000 [3]. In a study of OUD-hospitalization mortality, Whites, ages 50–64, Medicare beneficiaries with disabilities, and residents of lower-income areas were noted to have higher odds of opioid/heroin poisoning [26]. These studies provide one potential reason for higher mortality in Whites and are consistent with our observation of an independent association of White race with higher mortality during OUD-hospitalization, adjusted for age, sex, insurance, income, comorbidity, hospital region (rural/urban) and teaching status, location or bed size. Future studies are needed to assess the other underlying causes for higher mortality in Whites with OUD-hospitalizations.

Our observation of the association of male sex with higher inpatient mortality of OUD-hospitalization extends similar observations in people who underwent elective total joint replacement [27] or with pharmaceutical opioid related overdose deaths [28]. We also noted differences by region and by hospital characteristics in these outcomes, which extend similar findings for opioid and heroin-related overdose hospitalizations [8] to OUD-related hospitalizations.”

Some comorbidity indices have been developed to measure and weigh the overall burden of comorbidities, for example, Charlson comorbidity index(CCI), Deyo CCI, Romano CCI. Is there any special reason for choosing Deyo CCI in this article.

Response: All versions of Charlson index have been validated, we agree whole-heartedly with the reviewer. We chose the Deyo-Charlson version as one of the commonly used versions of this comorbidity index.

Minor point: Table 4: The form after “Rural” is empty. Should it be “Ref”?

Response: Thank you for catching this error, we have fixed it.

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

Submitted filename: Response_Reviewers_R2_OUD.docx

Click here for additional data file.

3 Feb 2020

National U.S. Time-trends in Opioid Use Disorder Hospitalizations and Associated Healthcare Utilization and Mortality

PONE-D-19-24166R1

Dear Dr. Singh,

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,

Sandra C. Buttigieg, MD PhD FFPH MSc MBA MMCFD

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 #1: All comments have been addressed

Reviewer #2: 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 #1: Yes

Reviewer #2: Yes

**********

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

Reviewer #1: Yes

Reviewer #2: 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 #1: Yes

Reviewer #2: 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 #1: Yes

Reviewer #2: 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 #1: I have reviewed the revised version and believe the authors have satisfactorily addressed the comments of the reviewers. My only remaining concern is that when considering hospitalization data in a life table way, one would usually remove those individuals who have a primary secondary diagnosis of cancer as often an opioid overuse/abuse

may be due to end stage cancer and pain relief which is considered acceptable in most settings, In other studies this is acknowledged and those individuals are removed prior to the analysis and/or are enumerated in the preliminary methods portion of the paper, I did not see this and so will leave it up to the editors to determine if they wish the authors to further comment on this adjustment.

Reviewer #2: The author's responce is complete, espeically in the part of first reviewer. I think expanded discussion and additional result will make this article meaningful.

**********

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 #1: No

Reviewer #2: Yes: Hsien-Yuan, Chang

7 Feb 2020

PONE-D-19-24166R1

National U.S. time-trends in opioid use disorder hospitalizations and associated healthcare utilization and mortality

Dear Dr. Singh:

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

Professor Sandra C. Buttigieg

Academic Editor

PLOS ONE