Demographics of patients receiving electroconvulsive therapy based on state-mandated reporting data

Abstract

Objectives:

Electroconvulsive therapy (ECT) is an effective treatment for depression and other psychiatric conditions. There is little comprehensive data on how many patients receive ECT in the United States, nor about the demographics of ECT recipients. This study characterizes the demographics of those receiving ECT, and how these demographics may have changed with time.

Methods:

Freedom of information requests for all data from record keeping inception through January 2019 were sent to the Department of Health or equivalent agency of states which mandate reporting of ECT. Information about demographics and the number of treating facilities were extracted.

Results:

Data was obtained on 62,602 patients receiving treatment in three states (California, Illinois, Vermont). Overall, 62.3% were women. Fewer than 1% of patients treated were under age 18, while 30.3% were 65 or older. White patients received a disproportionate proportion of treatments, with all other races underrepresented. The total number of facilities offering ECT in the three states declined over the study period.

Conclusions:

Recipients of ECT are more likely to be female, more likely to be elderly, and more likely to be white than the average person in their state.

Introduction

Electroconvulsive therapy (ECT) is an effective treatment for both unipolar and bipolar depression,^1,2 and is likely cost effective after as few as two failed medication trials.³ Despite their demonstrated efficacy over more than 80 years of clinical use, ECT devices were only FDA approved in 2018 for depression and catatonia,⁴ and their use is restricted by law in many states.⁵ These specific restrictions include limitations on patient age, number of annual treatments, and specific consent language, among others. Additionally, some states require at least annual reporting by facilities administering ECT to the Department of Health or equivalent state agency, generally including information about the number of procedures administered, the demographics of patients, and any reported complications. As there is no national database of ECT treatments or patients, these statewide reports are likely the largest and most comprehensive source of information about the use of ECT services. This study analyzes the demographics of ECT recipients based on this state-mandated reporting data.

Methods

Requests were sent to the Department of Health or equivalent agency of California, Illinois, and California from April through August 2019. Data was requested for all records held by the state agency from inception of record keeping through January 1^st 2019. Data from Texas have been previously reported,^6–8 and are not re-reported here. Information on the number of patients, patient demographics, and number of treating facilities was extracted. Data from California was obtained on a per-hospital basis, rather than statewide; as a result, information about rare events (less than 10 patients in a category) was redacted by the state. Redacted data were extrapolated as the total number of redacted patients in the dataset divided by the number of hospitals whose data was redacted. Statewide demographic data is taken from the 2010 American Community Survey 1-year estimates for the years 2001-2018.⁹ Differences in categorical variables were assessed using chi-square tests. Statistics were calculated using Prism (v 8.2.1, San Diego, CA).

Results

California, Illinois, and Vermont, together represent a population of 50.7 million individuals in 2010 (16.5% of the US population). There was variation in records retention among the three states. California data was retained for a 10 year period on a per-hospital basis, as described above. Illinois data was reported quarterly from hospitals and aggregated into annual data by the state. The state was only able to provide records for a subset of the requested period, and indicated that they were unable to locate the remaining data. Vermont maintains per-patient results, and for privacy reasons released annual data broken down per hospital. In total 62,602 patients received ECT in the three states, with women representing 62.3% of patients and 50.4% of the overall state population (Table 1). The rate of administration of ECT varied among states and years (Figure 1), from a high of 2.4 patients/10,000 residents in Vermont in 2014 to a low of 0.5 per 10,000 residents in California in 2015. The proportion of female patients remained relatively constant in California and Illinois, with greater variability in Vermont (Figure 2). Age breakdowns were available from all three states, with patients in their fifth decade and beyond making up 73.5% of treatments (Table 2). In total people under age 18 make up 24.7% of the three states, and receive fewer than 1% of treatments. People over age 65 represent 11.7% of the state populations, but make up 30.3% ECT recipients.

Table 1:

total number of patients and gender breakdown for each of the three states during the listed reporting period. In total, 62,602 patients were treated with ECT during the years for which data are available. There is an approximately 2:1 ration of female to male patients.

State	Years	Total Pts	Male	%	Female	%	Not stated	%
CA	2008-2017	39,114	14,936	38.2%	24,178	61.8%	0	0.0%
IL	2006-7; 2013-18	21,494	7,132	33.2%	13,726	63.9%	636	3.0%
VT	2001-2018	1,994	899	45.1%	1,095	54.9%	0	0.0%
Total:		62,602	22,967	36.7%	38,999	62.3%	636	1.0%

Figure 1:

Rates of ECT administration by state per year. The rate of ECT administration varies among states, with Vermont treating a higher fraction of residents in 2018 than in 2001, and California declining in treatment numbers between 2008 and 2017.

Figure 2:

Fraction of female patients per year, separated by state. There is a steady female predominance in California and Illinois throughout the study period, with Vermont showing greater variability in gender ratio.

Table 2:

age distribution of patients receiving ECT. Overall 30.3% of patients were age 65 or greater, while less than 1% were under age 18. Nearly three quarters of treatments were given to patients in at least their fifth decade of life.

State	Years	Ages
CA	2008-2017	0-17		18-24		25-44		45-64				65+		Missing
		247	0.6%	2,035	5.2%	8,211	21.0%	16,493	42.2%			12,065	30.8%	62	0.2%
IL	2006-7; 2013-18			18-21		22-40		41-64
		21	0.1%	412	1.9%	3,871	18.0%	9,618	44.7%			6,597	30.7%	975	4.5%
VT	2009-2018			18-21		22-34		35-49		50-64	%
		1	0.1%	5	0.4%	132	10.4%	396	31.2%	457	36.0%	280	22.0%	0	0.0%

Data on race was available from California and Illinois. The racial distribution of ECT recipients differs significantly from the overall racial distribution of each state (California p < 0.0001; Illinois p < 0.0001). White patients received more than 80% of treatments in both states, and are the only race to be overrepresented among treatment recipients relative to the population of the state (Table 3). All other races were less represented among treatment recipients then in the population, a trend that remained consistent in time even accounting for changes in demographics of the state during the study period (Figure 3).

Table 3:

breakdown of ECT recipients by race compared to the demographics of the state as of the 2010 American Community Survey. White patients are overrepresented among ECT recipients, while all other races received treatments at a lower rate than their share of the population.

State	Years		White	%	Hispanic	%	Black	%	Asian	%	Native American	%	Other/Unk	%	Missing	%
CA	2008-2017	ECT Recipients	31438	80.4%	3145	8.0%	1332	3.4%	1920	4.9%	128	0.3%	1151	2.9%	0	0.0%
		State overall		40.0%		37.7%		5.8%		12.9%		0.8%		2.8%
IL	2006-7; 2013-18	ECT Recipients	18245	84.9%	851	4.0%	964	4.5%	218	1.0%	62	0.3%	379	1.8%	775	3.6%
		State overall		63.6%		15.9%		14.3%		4.6%		0.1%		1.5%

Figure 3:

Number of patients per 10,000 residents per year broken down by race in Illinois (top) and California (bottom). Both states show significant overrepresentation of white patients relative to all other races.

During the period reflected in reports, there is an overall decrease in the number of hospitals offering ECT (Figure 4). Illinois had 52 facilities performing ECT in 2001, dropping to 12 facilities in 2018. Likewise, California went from having 52 facilities in 2008 to 38 in 2017.

Figure 4:

Number of facilities reporting ECT treatments by state and year. Both California and Illinois have a sharp decline in the number of facilities reporting treatments during the study period.

Discussion

This study reports on data from three US states during a period of up to 17 years, representing 62,602 patients. Our results indicate that the average ECT recipient is more likely to be female (1.65:1 F:M), more likely to be elderly (30.3% over age 65 vs. 11.7% of population), and more likely to be white (83% white recipients vs. 48.2% of population) than the average person in their state. These trends persisted throughout the study period, encompassing 2001-2018 (with variable date ranges for different states).

Data from California have been reported previously in the periods of 1977-1983 and 1984-1994.^10,11 In the former period, an average of 2,661 patients received ECT per year, of whom 69% were female, 91.7% were white, and 37.0% were age 65 or older, similar to the proportions receiving ECT in our data from 2008–2017. Likewise, reports from Texas indicate similar gender, race, and age proportions as seen in the states studied in this paper.^6–8

ECT is predominantly a treatment for depressive disorders, and major depressive disorder is approximately twice as common in women than in men,^12,13 which may explain the gender imbalance seen in our data. It is also possible that the difference is attributable to physicians differentially offering treatments to patients on the basis of gender, with one study suggesting that women are offered ECT sooner than men.¹⁴ Likewise, male and female patients may have different preferences for treatment with ECT that affect their willingness to undergo the procedure. Furthermore, men and women may have different levels of social supports, with divorced men in particular having less connection to and support from family than divorced women.¹⁵ This, combined with the differences in the way the symptoms of depression manifest themselves across the sexes may make it harder to recognize depression in men and seek appropriate treatment.

The age disparity, with elderly patients receiving a larger share of treatments, may be in part due to the extensive research supporting the effectiveness of the technique in depressed elders,¹⁶ although some research suggests equal effectiveness throughout the age range¹⁷ and practice guidelines also support the use of ECT in adolescents.¹⁸ Nonetheless, less than 1% of those treated in our dataset were under age 18, perhaps reflecting an under-treatment of this population who might benefit from ECT. Some of this under-treatment may be an effect of legislation that restricts adolescents’ access to ECT. Laws in California ban ECT for those younger than 12, and patients between 12 and 15 can only be treated in emergent situations and with the approval of three separate child psychiatrists, none of whom can be the patient’s own physician. Likewise in Illinois, ECT for patients under age 18 requires parental or guardian consent, approval by a court, and the agreement of two licensed psychiatrists. Vermont, in contrast, has no statues restricting the use of ECT in minors.⁵

Overrepresentation of geriatric patients among ECT recipients relative to their prevalence in the population may reflect several factors including greater confidence in diagnosis and efficacy in this age group, greater difficulty tolerating medications that reflects well known differences in volume of distribution and efficiency of drug metabolism between the elderly and younger adult patients, more co-morbid medical conditions requiring their own medications, and greater clinical urgency in treating severe psychiatric disorders in patients with likely less functional reserve, or a combination of other factors.¹⁹ This pattern of increased utilization of ECT in geriatric patients has been found in previous studies of utilization in the 1980s, when 34% of ECT recipients were aged 65 or older.²⁰

The large racial predominance of white patients relative to all others and has been noted in prior ECT epidemiologic studies.^7,8,21–23 As treatment outcomes appear similar regardless of the race of the patient,²⁴ this difference in utilization rates warrants further study. Multiple potential causes for this disparity have been proposed including racial differences in diagnosis of mood vs. psychotic disorders,²⁵ cultural differences in overall mental health utilization,²⁶ differences in preference among treatment options,^27,28 socioeconomic variables that affect resource availability in predominantly non-white communities, and confounding by other variables including age and sex, but none of these fully explain the differences in treatment rates.²³

One strength of our study is the use of state-mandated reporting data that captures outcomes regardless of payment source or treatment facility. This is in contrast to studies focusing in insurance databases that only capture privately insured Americans,²⁹ data from inpatient admissions which excludes recipients of outpatient ECT,³⁰ or data from Veteran’s Administration hospitals which are only accessible by a specific subpopulation.²¹ Given the variability of ECT use among different regions of the United States,³¹ as well as significant variation in legal regulation of ECT,⁵ it is difficult to extrapolate from the states in this study to produce an overall estimate of the number of patients treated with ECT nationwide.³² Previous estimates using practice surveys of psychiatrists calculated a national rate of 4.9 patients per 10,000 population in 1988-1989, with wide variation among regions studies (from 0.4 to 81.2 per 10,000).³¹ Likewise, meta-analyses of US and international data similarly show wide variations in ECT utilization, with a cumulative estimate of 1.7 per 10,000 patients worldwide,³³ with national rates of between 0.11 per 10,000 in Poland to 4.4 per 10,000 in Victoria, Australia.³⁴ No state in our sample had ECT utilization rates greater than 2.4 per 10,000 in any year, less than half of the 1988-1989 estimate. Whether that is due to comparatively low use of the procedure in the states studied here or a change in the nationwide rate of ECT is unknown.

Limitations of our study include its retrospective observational nature and our intrinsic reliance on summary data from secondary sources. As with any database study, there may be unintentional errors in reporting, however given the legal mandate to report treatments in these states it is unlikely that underreporting would be systematic in nature or biased with respect to demographics of patients. No state mandates reporting from federal facilities (although VA data is sometimes reported to Vermont), which prevents even these state databases from being truly comprehensive. Furthermore, reports are made by individual facilities, and thus each may have differing methods to determine demographic information such as race and gender which may hinder accurate reporting of multi-racial individuals or those of Hispanic background. Together these issues limit our ability to determine causation for the trends observed here. Nonetheless, these data suggest that inadequate treatment of mood disorders in older minority patients is an issue that deserves further consideration.

In conclusion, this study describes the demographics of patients treated with electroconvulsive therapy in the states of California, Illinois, and Vermont. There is nearly a 2:1 ratio of women to men, and patients over age 65 make up 30.3% of those treated. White patients are overrepresented relative to other races. Further studies are needed to investigate the public health and clinical implications of these disparities among ECT recipients.