Of the total 5,331,843 ED visits during 2017–2019 in Maryland, frequent visits made up 24.40% (1,301,301) of all visits while frequent users made up 7.16% (234,973) of all ED users. For each of the following variables, missing observations were dropped: 0.0022% of all observations for age; 0.014% for race; 0.0037% for location; 0.63% for income; 4.77% for distance; 2.06% for charge. In 2017, 7.43% were frequent ED users, compared to 7.18% in 2018 and 6.88% in 2019. Among the frequent users, there were 19,160 (10.22%) that reported recurring ED visits in 2017 and 2018 and 9,177 (4.90%) in 2017, 2018, and 2019. Figure 1 presents a chart depicting the top 10 ICD-10 diagnoses that account for frequent visits in the ED in Maryland during 2017–2019. The top ten primary diagnoses included chest pain (2.71%, 35,251), other chest pain (2.34%, 30,371), abdominal pain (1.86%, 24,249), upper respiratory infection (1.84%, 23,895), Urinary tract infection (1.64%, 21,394), headache (1.59%, 20,649), asthma (1.38%, 17,959), low back pain (1.15%, 14,917), nausea with vomiting (1.05%, 13,667) and viral infection (0.93%, 12,146).
The total charges associated with all frequent visits in Maryland during 2017–2019 were $1.61 billion and the top ten diagnoses had total charges of $282 million. Table 1 breaks down the average and total charges associated with the top ten diagnoses of frequent ED visits identified earlier. With an average charge of $1,941 [95% CI, 1,920-1,962] and $2,039 [95% CI, 2,016 − 2,061] respectively, chest pain and other chest pain diagnoses were the top two diagnoses that collectively made up $130 million of the total charges. While charges associated with frequent abdominal pain visits made up $30.33 million [95% CI 29.89–30.77] of the total charges, urinary tract infection ($26.16 mil, 95% CI 25.76–26.57]), headache ($20.67 mil 95% CI 20.35–20.98) and unspecified asthma with acute exacerbation ($ 20.48 mil 95% CI 20.09–20.87) collectively exceeded $40 million. Lastly, nausea with vomiting (17.92 mil 95% CI 17.60-18.23), acute upper respiratory infection (15.22 mil 95% CI 15.04–15.40) and viral infection (9.32 mil 95% CI 9.18–9.46) contributed to $42.46 million.
Bivariate Analysis:
Frequent and non-frequent ED users
Of the total 3,279,651 ED users, we estimated 234,973 (7.16%, 95% CI 7.14–7.19) frequent ED users and 3,044,678 (92.84%, 95% CI 92.81–92.86) non-frequent ED users during 2017–2019 (Table 2). The distribution of frequent and non-frequent ED users was statistically significant by sex, age, and race. Female patients accounted for 59.88% [95% CI, 59.69–60.08] of all frequent ED users, in contrast with 54.71% [95% CI, 54.65–54.77] of non-frequent users. Majority of the frequent ED users (42.86%, 95% CI 42.66–43.06) were between 21 and 44 years of age compared to 32.98% [95% CI, 32.93–33.03] for non-frequent users. While 54.16% [95% CI, 53.96–54.36] of frequent users were NH Black in contrast with 39.73% [95% CI, 39.67–39.78] of NH Black non-frequent users.
Table 2
Patient and community level characteristics of frequent and non-frequent ED users, 2017–2019
PATIENT AND COMMUNITY LEVEL CHARACTERISTICS | Non-frequent ED users [%, 95%CI] | Frequent ED users [%, 95%CI] | P-value* |
Total | 92.84 [92.81–92.86] (3,044,678) | 7.16 [7.14–7.19] (234,973) | |
Sex | | | < 0.000 |
Male | 45.29 [45.23–45.34] | 40.12 [39.92–40.32] | |
Female | 54.71 [54.65–54.77] | 59.88 [59.69–60.08] | |
Age group | | | < 0.000 |
0–20 years | 25.77 [25.72–25.82] | 18.58 [18.42–18.73] | |
21–44 years | 32.98 [32.93–33.03] | 42.86 [42.66–43.06] | |
45–64 years | 24.26 [24.21–24.30] | 25.02 [24.84–25.19] | |
65 and over | 16.99 [16.85–17.03] | 13.55 [13.41–13.69] | |
Race | | | < 0.000 |
NH White | 44.74 [44.68–44.79] | 37.10 [36.91–37.30] | |
NH Black | 39.73 [39.67–39.78] | 54.16 [53.96–54.36] | |
Hispanic | 9.41 [9.37–9.44] | 5.44 [5.35–5.54] | |
NH Other | 6.13 [6.10–6.16] | 3.29 [3.22–3.37] | |
Payer | | | |
Private | 38.67 [38.61–38.72] | 18.11 [17.97–18.25] | < 0.000 |
Medicaid | 29.33 [29.29–29.38] | 51.98 [51.79–52.17] | < 0.000 |
Medicare | 22.07 [22.02–22.11] | 23.67 [23.51–23.84] | < 0.000 |
Other | 9.93 [9.89–9.96] | 6.23 [6.16–6.30] | < 0.000 |
Location | | | |
Large central and fringe metros | 86.14 [86.10-86.17] | 83.25 [83.10-83.39] | < 0.000 |
Medium and small metros | 11.02 [10.99–11.06] | 13.00 [12.87–13.13] | < 0.000 |
Micropolitan areas | 0.99 [0.98–1.01] | 1.67 [1.62–1.73] | < 0.000 |
Non-core areas | 1.66 [1.65–1.67] | 2.08 [2.02–2.13] | < 0.000 |
GRPCI (Grouped Charlson Comorbidity Index) | | | |
No chronic condition | 77.21 [77.16–77.25] | 66.77 [66.63–66.93] | < 0.000 |
One chronic condition | 15.91 [15.88–15.95] | 21.40 [21.29–21.52] | < 0.000 |
Two or more chronic condition | 6.88 [6.85–6.90] | 11.82 [11.71–11.92] | < 0.000 |
Median Household income** | | | |
First Quartile | 33.53 [33.48–33.58] | 48.16 [47.97–48.35] | < 0.000 |
Second Quartile | 26.41 [26.36–26.46] | 24.03 [23.87–24.18] | < 0.000 |
Third Quartile | 22.32 [22.27–22.36] | 18.01 [17.87–18.15] | < 0.000 |
Fourth Quartile | 17.74 [17.69–17.78] | 9.80 [9.69–9.91] | < 0.000 |
Mean Distance traveled | 6.84 miles [6.83–6.85] | 6.15 miles [6.13–6.18] | < 0.000 |
Average Charge | $1,164.81 [1,163.22-1,166.41] | $1,208.68 [1,205.21-1,212.15] | < 0.000 |
Weekend | | | |
Yes | 27.30 [27.26–27.34] | 26.38 [26.32–26.45] | < 0.000 |
No | 72.69 [72.66–72.74] | 73.61 [73.55–73.68] | < 0.000 |
*- Chi-square test was run for constant categorical variables [sex, age-group and race] and t-tests were run for the proportions estimated for each categories of categorical variables that were averaged over the study period. |
** In 2017, 2018 and 2019, the first quartile corresponded to income below $60,999, $66,1999 and $69,999, respectively. In 2017, the second quartile corresponded to income between $61,000 and $76,499 66,200-$81,299 in 2018 and $70,000–85,299 in 2019. The third quartile corresponded to income between $76,500 and $98,499 in 2017, $81,300-$103,299 in 2018, and $85,300-$110,099 in 2019. Fourth quartile corresponded to income above $98,500 in 2018, $103,300 in 2018, and $110,100 in 2019. |
Over half of the frequent users were insured by Medicaid (51.98%, 95% CI 51.79–52.17), as opposed to 29.33% [95% CI, 29.29–29.38] of non-frequent users. Non-frequent ED users were mainly insured by private insurance (38.67%, 95% CI 38.61–38.72). In contrast with 83.25% [95% CI, 83.10-83.39] of frequent users, 86.14% [95% CI, 86.10-86.17] of non-frequent users lived in the large central and fringe metropolitan area.
A higher share of frequent ED users had one or more chronic conditions than non-frequent users (33.22% compared to 22.79%). While 48.16% [95% CI, 47.97–48.35] of frequent users belonged to the lowest income quartile of the estimated median household income for Maryland, only 33.53% [95% CI, 33.48–33.58] of non-frequent users belonged to the same quartile. Frequent users (6.15 miles, 95% CI 6.13–6.18) lived closer to the hospital than non-frequent users (6.84 miles, 95% CI 6.83–6.85). The average charges for frequent and non-frequent users reported a statistically significant difference ($1,208.68 and $1,164.81 in current US dollars respectively). Over 72% of the frequent and non-frequent users used the ED on a weekday.
Recurrent and non-recurrent ED users in 2017–2018 and 2017–2019
Among the 187,391 unique frequent users, there were 19,160 (10.22%) that visited the ED in two subsequent years (time period 1: 2017–2018) and 9,177 (4.90%) during all three years, (time period 2: 2017–2019). The distribution of recurrent and non-recurrent frequent ED users differed by sex, age, and race at 95% l.o.s (Table 3). As compared to 36.34% [95% CI 35.66–37.02] male patients, 63.66% [95% CI, 62.98–64.34] females visited the ED frequently in time period 1. Consistently, while 34.64% [95% CI, 33.67–35.62] male frequent users visited the ED in time period 2, 65.36% [95% CI, 64.38–66.33] female frequent users were recurring patients during all three years. When compared across age and racial groups, 47.95% [95% CI, 47.25–48.66] of recurrent patients in time period 1 were between 21–44 years, while 41.06% [95% CI, 40.82–41.29] were non-recurrent frequent users. Almost half (48.36%, 95% CI 47.33–49.38) of the recurrent frequent ED users during time period 2 were between 21–44 ages (compared to 41.42%, 95% CI 41.19–41.65 for non-recurrent frequent users). Over half, 57.54% [95% CI, 56.84–58.24] of recurrent frequent ED users in 2017 and 2018 were Black, compared to 52.62% [95% CI, 52.38–52.86] for non-recurrent frequent users. Similar data were reported for recurring visits in 2017, 2018 and 2019 (59.03%, 95% CI 58.02–60.03, and 52.82%, 95% CI 52.58–53.05, for recurrent and non-recurrent users respectively).
Table 3
Characteristics of recurrent and non-recurrent frequent users in Maryland, SEDD 2017–2019.
Visits in 2017 [n = 191,212] | Visits in 2018 [%, 95%CI] | P-value | Visits in 2018 and 2019 [%, 95%CI] | P-value |
% of patients [No.] | Non-recurrent | Recurrent | | Non-recurrent | Recurrent | |
Total | 89.78 [89.64–89.91] 168,231 | 10.22 [10.09–10.36] 19,160 | | 95.10 [95.00-95.19] 178,214 | 4.90 [4.80–4.99] 9,177 | |
Sex | | | < 0.000 | | | < 0.000 |
Male | 41.41 [41.17–41.64] | 36.34 [35.66–37.02] | | 41.21 [40.98–41.44] | 34.64 [33.67–35.62] | |
Female | 58.59 [58.36–58.83] | 63.66 [62.98–64.34] | | 58.79 [58.56–59.02] | 65.36 [64.38–66.33] | |
Age group | | | < 0.000 | | | < 0.000 |
0–20 years | 20.81 [20.61-21.00] | 12.08 [11.63–12.55] | | 20.44 [20.25–20.63] | 9.72 [9.13–10.34] | |
21–44 years | 41.06 [40.82–41.29] | 47.95 [47.25–48.66] | | 41.42 [41.19–41.65] | 48.36 [47.33–49.38] | |
45–64 years | 23.37 [23.17–23.58] | 29.68 [29.04–30.33] | | 23.61 [23.41–23.80] | 31.99 [31.05–32.95] | |
65 and over | 14.76 [14.59–14.93] | 10.28 [9.86–10.72] | | 14.53 [14.37–14.70] | 9.93 [9.33–10.56] | |
Race | | | < 0.000 | | | < 0.000 |
NH White | 37.55 [37.32–37.78] | 36.88 [36.19–37.56] | | 37.55 [37.32–37.77] | 36.23 [35.25–37.22] | |
NH Black | 52.62 [52.38–52.86] | 57.54 [56.84–58.24] | | 52.82 [52.58–53.05] | 59.03 [58.02–60.03] | |
Hispanic | 6.15 [6.03–6.26] | 3.24 [2.99–3.50] | | 6.01 [5.90–6.13] | 2.64 [2.33–2.98] | |
NH Other | 3.68 [3.59–3.77] | 2.34 [2.14–2.57] | | 3.62 [3.53–3.70] | 2.10 [1.83–2.42] | |
Payer | | | | | | |
Private | 19.94 [19.77–20.12] | 13.26 [12.84–13.67] | < 0.000 | 19.66 [19.50-19.83] | 11.40 [10.84–11.96] | < 0.000 |
Medicaid | 49.83 [49.61–50.05] | 57.49 [56.86–58.12] | < 0.000 | 50.20 [49.99–50.42] | 58.45 [57.54–59.37] | < 0.000 |
Medicare | 23.35 [23.16–23.54] | 24.84 [24.26–25.42] | < 0.000 | 23.35 [23.16–23.54] | 26.50 [25.64–27.35] | < 0.000 |
Self-pay/No-charge | 6.87 [6.79–6.97] | 4.41 [4.22–4.59] | < 0.000 | 6.79 [6.70–6.87] | 3.65 [3.42–3.89] | < 0.000 |
Location | | | | | | |
Large central and fringe metros | 83.52 [83.35–83.69] | 82.10 [81.57–82.63] | < 0.000 | 83.43 [83.26–83.60] | 82.27 [80.51–83.03] | < 0.01 |
Medium and small metros | 12.70 [12.54–12.85] | 14.02 [13.54–14.50] | < 0.000 | 12.77 [12.62–12.92] | 14.09 [13.40-14.78] | < 0.01 |
Micropolitan areas | 1.71 [1.65–1.77] | 1.68 [1.50–1.85] | 0.7165 | 1.72 [1.66–1.77] | 1.52 [1.29–1.76] | 0.1514 |
Non-core areas | 2.07 [2.00-2.14] | 2.21 [2.00-2.41] | 0.1970 | 2.08 [2.02–2.15] | 2.12 [1.83–2.40] | 0.8376 |
GRPCI | | | | | | |
No chronic condition | 68.97 [68.80-69.15] | 60.57 [60.05–61.08] | < 0.000 | 68.66 [68.49–68.83] | 57.56 [56.81–58.30] | < 0.000 |
One chronic condition | 20.04 [19.90-20.17] | 25.23 [24.83–25.63] | < 0.000 | 20.26 [20.13–20.39] | 26.55 [25.98–27.12] | < 0.000 |
Two or more chronic condition | 10.99 [10.87–11.11] | 14.21 [13.83–14.58] | < 0.000 | 11.08 [10.97–11.20] | 15.89 [15.33–16.45] | < 0.000 |
Median Household income | | | | | | |
First Quartile | 46.70 [46.48–46.93] | 51.33 [50.69–51.97] | < 0.000 | 46.89 [46.68–47.11] | 52.76 [51.84–53.68] | < 0.000 |
Second Quartile | 24.22 [24.04–24.41] | 24.03 [23.52–24.53] | 0.4959 | 24.23 [24.05–24.41] | 23.65 [22.94–24.37] | 0.1574 |
Third Quartile | 18.49 [18.32–18.66] | 16.70 [16.25–17.15] | < 0.000 | 18.42 [18.25–18.58] | 16.21 [15.58–16.84] | < 0.000 |
Fourth Quartile | 10.58 [10.44–10.71] | 7.94 [7.61–8.28] | < 0.000 | 10.46 [10.33–10.60] | 7.38 [6.91–7.84] | < 0.000 |
Weekend | | | | | | |
Yes | 26.36 [26.27–26.44] | 26.50 [26.33–26.67] | 0.296 | 26.36 [26.28–26.45] | 26.56 [26.34–26.78] | 0.2994 |
No | 73.64 [73.55–73.73] | 73.50 [73.33–73.67] | 0.296 | 73.64 [73.55–73.72] | 73.44 [73.22–73.66] | 0.2994 |
Average Charge | $1,201.59 [1,197.24-1,205.93] | $1218.03 [1,208.05-1,228.01] | 0.015 | $1,200.10 [1,195.93-1,204.28] | $1,264.79 [1,250.77-1,278.80] | < 0.000 |
Average Distance | 6.30 miles [6.27–6.33] | 5.80 miles [5.73–5.88] | < 0.000 | 6.28 miles [6.24–6.31] | 5.67 miles [5.57–5.77] | < 0.000 |
*- Chi-square test was run for constant categorical variables [sex, age-group and race] and t-tests were run for the proportions estimated for each categories of categorical variables that were averaged over the study period. |
Majority of the recurrent frequent users were insured by Medicaid. Over half (57.49%, 95% CI 56.86–58.12) of recurrent frequent users in 2017 and 2018 reported to have Medicaid, compared to 49.83% [95% CI, 49.61–50.05] for non-recurrent frequent users. Consistently, Medicaid patients accounted for 58.45% [95% CI, 57.54–59.37] of recurrent users in 2017, 2018 and 2019. Residents of large central and fringe metro areas accounted for 82.27% [95% CI, 80.51–83.03] of recurring frequent users in 2017, 2018 and 2019 (compared to 83.43%, 95% CI 83.26–83.60, for non-recurring users). Similar results were found for recurrent frequent users in 2017 and 2018. Over a quarter (25.22%, 95% CI 24.83–25.63) of recurrent frequent users in 2017 and 2018 had a single chronic condition (20.04%, 95% CI 19.90-20.17, for non-recurring patients). Similar shares were reported for recurrent patients in all three years (26.55%, 95% CI 25.98–27.12, and 20.26%, 95% CI 20.13–20.39). Most recurrent frequent users belong to the lowest income quartile (51.33%, 95% CI 50.69–51.97, in time period 1, and 52.76%, 95% CI 51.84–53.68, in time period 2). The average visit charge for a recurrent frequent user $1,264.79 [95% CI, 1,250.77-1,278.80] in time period 2 was higher (and statistically significant) from that of a non-recurring frequent user ($1200.10, 95% CI 1,195.93-1,204.28). A recurrent frequent user traveled 0.5 and 0.6 miles lesser than a non-recurrent frequent user in period 1 and 2 respectively.
Multivariable Analysis
Results of the multivariate logistic regressions are presented in Table 4. Model 1 compared frequent ED users to non-frequent ED users, while Models 2 and 3 compared recurrent frequent ED users to non-recurrent users in 2017–2018 and 2017–2019 respectively.
Table 4
Logistic Regression Results, Factors associated with frequent and recurrent ED users, 2017–2019.
| Model 1: All frequent users | Model 2: Recurrent frequent users in 2017 and 2018 | Model 3: Recurrent frequent users in 2017, 2018 and 2019. |
Comparison Group: | Non-frequent user | Non-recurrent frequent user | Non-recurrent frequent user |
Total Sample Size (n) | 3,279,651 | 187,391 | 187,391 |
Sex | | | |
Male | Reference | Reference | Reference |
Female | 1.16 (1.15, 1.17) | 1.22 (1.18, 1.25) | 1.30 (1.24, 1.36) |
Age group | | | |
0–20 years | Reference | Reference | Reference |
21–44 years | 2.04 (2.02, 2.07) | 1.99 (1.90, 2.09) | 2.41 (2.24, 2.60) |
45–64 years | 1.49 (1.47, 1.52) | 1.79 (1.69, 1.89) | 2.19 (2.01–2.38) |
65 and over | 0.79 (0.78, 0.81) | 0.75 (0.69, 0.82) | 0.80 (0.71, 0.90) |
Race | | | |
NH White | Reference | Reference | Reference |
NH Black | 1.26 (1.25, 1.28) | 1.05 (1.01, 1.08) | 1.08 (1.03, 1.13) |
Hispanic | 0.70 (0.69, 0.71) | 0.65 (0.59, 0.71) | 0.60 (0.52, 0.68) |
NH Other | 0.69 (0.67, 0.70) | 0.74 (0.67, 0.82) | 0.72 (0.62, 0.84) |
Payer | | | |
Private | Reference | Reference | Reference |
Medicaid | 4.03 (3.98, 4.08) | 2.00 (1.90, 2.11) | 2.41 (2.24, 2.60) |
Medicare | 2.75 (2.71, 2.80) | 2.09 (1.96, 2.23) | 2.62 (2.39, 2.86) |
Others | 1.21 (1.18, 1.23) | 0.77 (0.69, 0.86) | 0.65 (0.55, 0.78) |
Location | | | |
Large central and fringe metros | Reference | Reference | Reference |
Medium and small metros | 1.06 (1.05, 1.07) | 1.09 (1.04, 1.14) | 1.07 (1.00, 1.14) |
Micropolitan areas | 1.43 (1.38, 1.48) | 0.91 (0.81–1.03) | 0.78 (0.65, 0.93) |
Non-core areas | 0.97 (0.94, 1.00) | 1.02 (0.92, 1.14) | 0.93 (0.80, 1.09) |
GRPCI | | | |
No chronic condition | Reference | Reference | Reference |
One chronic condition | 1.79 (1.77, 1.81) | 1.94 (1.85, 2.04) | 2.20 (2.06, 2.35) |
Two or more chronic condition | 2.73 (2.69, 2.78) | 1.95 (1.83, 2.08) | 2.34 (2.14, 2.54) |
Median Household income | | | |
Fourth Quartile | Reference | Reference | Reference |
Third Quartile | 1.29 (1.27, 1.31) | 1.15 (1.08, 1.24) | 1.19 (1.08, 1.32) |
Second Quartile | 1.33 (1.12, 1.36) | 1.25 (1.17, 1.34) | 1.31 (1.19, 1.43) |
First Quartile | 1.69 (1.66, 1.71) | 1.23 (1.15, 1.31) | 1.33 (1.21, 1.46) |
Weekend | | | |
No (0) | Reference | Reference | Reference |
Yes (1) | 1.02 (1.00, 1.02) | 0.90 (0.84, 0.96) | 0.87 (0.80, 0.94) |
Logged average distance | 1.09 (1.09, 1.10) | 1.01 (0.99–1.03) | 1.03 (1.00, 1.05) |
Year | | | |
2017 | Reference | -- | -- |
2018 | 0.96 (0.95, 0.97) | -- | -- |
2019 | 0.93 (0.92, 0.94) | -- | -- |
Note: Values in bold indicate results that are not statistically significant |
Compared to males, women were 1.16 times [95% CI 1.15,1.17] more likely to be frequent users. They were also 30% [95% CI 1.24,1.36] more likely to be recurrent frequent ED users than male frequent users in time period 2. While patients aged 21 to 44 were twice more likely to visit the ED frequently as compared to children and adolescents (0–20 years), they also were 2.41 [95% CI 2.24,2.60] times more likely to be recurrent frequent users more than younger frequent users. NH Black ED users were 26% [95% CI 1.25,1.28] more likely to be frequently seen by an ED healthcare worker than NH White patients. Compared to NH White frequent ED users, NH Black frequent ED users were 8% [95% CI 1.03,1.13] more likely to be seen recurrently through 2017 to 2019.
Patients insured by Medicaid were 4 times [95% CI 3.98,4.08] more likely to be frequent ED users, compared to privately insured patients. They were also 2.41 times more likely to be recurring ED users (95% CI 2.24–2.60) while Medicare patients were 2.75 times [95% CI 2.71,2.80] more likely to be recurrent frequent ED users. Compared to residents of large central and fringe metropolitan areas, patients from micropolitan areas were 43% [95% CI 1.38,1.48] more likely to be frequent ED users, however less likely to be recurrent frequent users in time period 2 [OR = 0.78, 95% CI 0.65,0.93]. A patient with two or more chronic conditions was 2.73 times [95% CI 2.69,2.78] more likely to be frequently seen at the ED than patients with no chronic condition. Consistently, these patients were also 2.34 times [95% CI 2.14,2.54) more likely to be recurrent ED users than those with no chronic condition. Patients who belonged to the lowest income quartile were 69% [95% CI 1.66,1.71] more likely to be frequent ED users than patients who belonged to the highest income quartile. They were also 33% [95% CI 1.21,1.46] more likely to be recurrent frequent patients, compared to frequent users of the highest income quartile. When compared to non-frequent users, frequent users were 9% [95% CI 1.09, 1.10] more likely to use the ED as the distance to the ED from their residence increased. Patients visited the ED in 2018 and 2019 less often than in 2017.