Abstract
Electroconvulsive therapy (ECT) utilization rates have varied substantially over time and by location [1–5]. Some of this volatility may have been due to differences in data collection, but Hermann et al. [6, 7] showed that other factors including the availability of psychiatrists, primary care physicians and hospital beds, the legislative requirements of the respective jurisdictions, and prescriber variables were also influential. Even within the same hospital, the prescribing practices of psychiatrists have shown substantial variation [8]. Notwithstanding these differences, ECT recipients tend to be older women with affective disorders [3, 9–11].
The present study utilized a centralized database of every ECT administration reported under statute by all ECT facilities (public and private) in Victoria, Australia, to describe the utilization of ECT, and the age, sex and diagnosis and of those treated. The statutory database was also matched with the Victorian Psychiatric Case Register (VPCR) [12] to enable an examination of those factors associated with ECT treatment in the public sector. To our knowledge, only two published studies have examined patient-based predictors of ECT treatment [1, 3].
Method
Victorian context
Victoria's 4.7 million inhabitants (comprising 25% of the Australian population) are serviced by a mixed public and private mental health system; the key features of which derive from Australia's National Mental Health Strategy [13, 14]. Public sector services are available at no charge, whereas consumers accessing private services incur fees that are partially reimbursed by the Commonwealth Government's Medicare scheme. Any difference between the rebate and the fee is met by the individual or their health insurance fund.
Only a psychiatrist can prescribe ECT in Victoria, the criteria for consent being specified by legislation. Patients able to give informed consent may receive ECT in the public or private sector without further opinion or independent review. Patients involuntarily detained (a status only permitted in the public sector) can have consent given on their behalf by the psychiatrist. In these instances the legislation does not require a second opinion or independent review, although policy guidelines [15] indicate that this should be sought. Electroconvulsive therapy can be performed on either an inpatient or outpatient basis in either sector.
Under Victorian legislation, all ECT facilities must be licensed. It is a requirement of the legislation, and a condition of the license, that all public and private facilities report each month (via a paper-based system) the details of every ECT administration performed. The information is reported to the Chief Psychiatrist appointed by the Victorian State Government.
Study period and measures
Descriptive analysis: public and private sector ECT
Aggregated statutory data for the 1998–1999 financial year were used to examine utilization rates and the descriptive characteristics of the ECT-treated group. The proportion of the mandatory monthly reports actually submitted during this period was 91.5%, and was similar for both public and private sectors.
A unique identifier (unit record number: URN) for the public system enabled non-duplicated enumeration of patients who received treatment at more than one service in this sector. However, each private hospital generated their own patient identifier. This introduced the potential for double-counting if a person received ECT in both public and private sectors, or in more than one private hospital. It was estimated by a process of matching date of birth, country of birth and sex that double-counting potentially accounted for no more than 6.6% of all URNs.
Crude age adjusted and age–sex specific utilization rates expressed as the number of people treated, and the number of administrations per 100 000 resident population per annum were calculated using mid-year 1999 population data estimated by the Australian Bureau of Statistics. For treated-person rates, a person was counted only once irrespective of the number of ECT administrations or courses received during the year. In calculating administration rates, ECT administrations forming part of a censored ‘course’ (i.e. spanning consecutive financial years), where a course was defined as ECT administrations separated by a period of not more than 2 days, were excluded from the analysis. The age, sex and diagnostic characteristics of those treated with ECT are reported as proportions of the overall ECT treated group. Age was at mid-year 1999. Diagnosis was the principal diagnosis for which ECT was given, as recorded at the time of each ECT administration. Where a person had more than one diagnosis recorded during a course of ECT (as occurred in 10.4% of ECT courses), or had received several courses of ECT during the 12-month period, the diagnosis assigned for the analysis was that of highest rank according to the following hierarchy: rank 1 (highest): schizophrenia; rank 2: schizoaffective disorder; rank 3: bipolar disorder (phase unspecified); rank 4: major depressive episode; rank 5: residual category (anxiety, somatoform and substance use disorders).
Multivariate analysis: public sector ECT
The VPCR was used to obtain further data on the ECT-treated group, and to identify an untreated comparison group for analyses of factors associated with ECT treatment in the public sector.
The VPCR is a purpose-built database extracted from the public mental health service on-line administrative data system, and contains linked information on age, sex, dates of episodes of care (both inpatient and outpatient), and the diagnosis for the episode of care. The VPCR contains records dating from 1961, and derives a ‘lifetime diagnosis’ based on the hierarchical allocation of diagnosis according to the following: rank 1 (highest) ‘organic’ disorders including dementia and other cognitive disorders; rank 2: schizophrenia and related psychoses; rank 3: major affective disorders; rank 4: residual category (anxiety and somatoform disorders, acute stress reactions, personality disorders, and alcohol and substance misuse).
A series of logistic regressions using the STATA software package, version 6 [16] were performed to examine the association between ECT treatment (the dependent variable) and the person's age, sex, and lifetime diagnosis.
The index group was limited to those ECT-treated patients in the public sector who had at least one admission during the 12-month sampling frame (n = 995; representing 70% of the total ECT public/ private treated group, and 96% of the public sector ECT-treated group). These were compared to a reference population of public mental health service patients admitted at least once during 1998–1999, who did not receive ECT, and who were aged 15 years or above (n = 12 741). Patients under 15 were excluded from the reference group, as the youngest ECT-treated patient was 15 years.
Regression results are reported as unadjusted and adjusted odds ratios, and are relative to the reference categories of being male, having an age of 15 years, and a lifetime diagnosis of the ‘residual category’ (rank 4).
Results
Utilization rates
In 1998–1999, 1535 patients received ECT in Victoria; accounting for 14 116 ECT administrations. This equated to a crude treated-person rate of 39.9 people per 100 000 resident population, and a crude administration rate of 330.3 ECT administrations per 100 000 resident population per annum. Age adjusted treated-person and administration rates were 44.0 persons, and 362.6 administrations, per 100 000 resident population per annum, respectively.
Table 1 demonstrates the considerable variation in rates (calculated per age and sex specific resident population) between age-sex specific groups. Utilization rates, measured as either persons treated, or ECTs administered, per 100 000 population per annum rose with increasing age (with the exception of the 85 + age group), and were higher (generally two-fold) for women in most age groups. Only in the 15–24-year-old group did men have a higher rate.
Age–sex specific rates per 100 000 age–sex specific resident population per annum
Treated group characteristics
Those treated ranged from 15 to 93 years; the proportion of the total number of ECT-treated people in each 10-year age bracket was similar except for at the extremes of age (Table 2).
Proportion of total treated persons in each 10-year age group
Overall, 62.8% of those treated were women. This proportion varied little according to age, ranging from 63.7% to 74.0% for each 10-year age group. The exception was the 15–24 age group in which women only represented 43.9%.
Depression was the principal diagnosis for which ECT was administered in 75.2% of those treated. An age–sex specific analysis indicated that the pre-eminence of affective disorder held irrespective of age or gender, except in 15–34-year-old males. In this group the principal diagnoses for which ECT was administered were the psychoses (Table 3).
Diagnostic proportions of the electroconvulsive therapy treated group, by age–sex group
Predictors of ECT
Of the 12 741 people aged 15 years and above admitted to a public mental health service at least once during 1998–1999, 7.8% (n = 995) were treated with ECT. Age, sex and diagnosis each independently predicted ECT treatment (Table 4).
Logistic regression model for electroconvulsive therapy in the public sector
Electroconvulsive therapy was almost twice as likely to have been administered to women as to men (independent of age or diagnosis), and each 10-year increase in age was associated with approximately a 25% increase in the likelihood of ECT treatment (independent of diagnosis or sex). Major affective disorders were associated with approximately 12 times increased likelihood of ECT compared to a lifetime diagnosis in the residual category of anxiety, somatoform, personality and substance use disorders. Patients with a lifetime diagnosis of schizophrenia were 4.5 times, and with an ‘organic’ disorder 2.6 times, as likely to receive ECT as those in the residual group. The increased likelihoods for various diagnoses were independent of age or sex.
Discussion
It is difficult to make meaningful comparisons of ECT utilization rates with those reported in the literature due to the marked intra- and interregional variation in utilization rates in the published literature, to differences between published studies and the present examination in data collection and reporting, and in the substantial variation in rates by age and sex demonstrated in the present study. Furthermore, it is not clear what would represent a significant – either statistical or clinical – difference between jurisdictions. Even within Victoria, it will be a challenge to determine whether any subsequent changes in utilization rates are of significance.
General trends in the characteristics of those treated, however, were consistent with those reported elsewhere. Consistent with other studies, the majority of ECT was given to women, especially of higher age, for an episode diagnosis of affective disorder [3, 9–11].
The greater proportion of ECT given for the psychoses with younger age has not been previously reported. It may represent a perceived greater responsiveness in younger age groups to non-ECT therapies for depression, or to a general reluctance to treat depression in young people with ECT (resulting in a lower proportion of ECT given for depression and correspondingly higher proportions given for other diagnoses). These hypotheses could not be tested in the present study.
The importance of age, female sex, and lifetime diagnoses of either major affective disorder or schizophrenia in predicting ECT treatment was comparable to other published reports, although effect sizes and the order of importance of factors differed. Thompson et al. [3], using data from 1975, 1980 and 1986 found that patients with mood disorders were twice as likely as those with schizophrenia (compared to almost three times in the present study), and that women were 1.42 times more likely than men (1.82 in the present study), to have been treated with ECT. Age was also an important predictor of ECT use. The differences between this report and the present study may have been due to annual variation (noted by Thompson et al. [1, 3]), to prescribing differences, or to the inclusion of fewer explanatory variables in the model developed for the present analysis. The inclusion of all variables utilized by Thompson et al. or of additional explanatory variables such as severity of illness, failure of other treatments, prior treatment with ECT, or consumer or prescriber preferences would probably reduce the estimated effect of each individual factor. This was unable to be undertaken in the present analysis.
Reasons for the greater likelihood of being treated with ECT in females couldn't be further explored with our available data. Possible explanations include a greater acceptance of ECT by women, or to a greater willingness to prescribe ECT to women.
Understanding the reasons underpinning the greater likelihood of ECT with age requires further research. The finding may have reflected reluctance to use ECT early in the course of an illness, to greater severity or treatment resistance with age, or to advantages of ECT over other therapies in treating of depression in the elderly [17].
Conclusions
Our data have shown ECT utilization in context with few legislative or resource constraints. While the results are in line with other reports, they raise questions concerning determination of an ‘appropriate’ rate of ECT utilization, and point to the possibile gender bias. Further research should identify whether or not there are timebased or regional variations in ECT utilization, and examine provider factors.
Footnotes
Acknowledgements
Thanks to Norman James, Ruth Vine and Jane Pirkis.