Abstract
There has been limited research on the quality of inpatient psychiatry, yet policies to expand access have increased, such as the use of Medicaid Section 1115 waivers for treatment in “Institutions for Mental Disease” (IMD). Using data from public records requests, we evaluated complaints, restraint, and seclusion from inpatient psychiatric facilities in Massachusetts occurring from 2008 to 2018, and compared differences in the rates of these events by IMD status. There were 17,962 total complaints, with 48.9% related to safety and 19.9% related to abuse (sexual, physical, verbal), and 92,670 episodes of restraint and seclusion. On average, for every 30 census days in a given facility, restraint, and seclusion occurred 7.47 and 1.81 times, respectively, and a complaint was filed 0.94 times. IMDs had 47.8%, 68.3%, 276.9%, 284.8%, 183.6%, and 236.1% greater rates of restraint, seclusion, overall complaints, substantiated complaints, safety-related complaints, and abuse-related complaints, respectively, compared to non-IMDs. This is the first known study to describe complaints from United States inpatient psychiatric facilities. Policies should strengthen the implementation of patients’ rights and patient-centeredness, as well as external critical-incident-reporting systems.
3–5 Key Points and Findings
There were 17,962 complaints and 92,670 episodes of restraint and seclusion from Massachusetts inpatient psychiatric facilities during 2008–2018
“Institutions for Mental Diseases” had much higher rates of complaints, restraint, and seclusion
This is the first known study to describe complaints from United States inpatient psychiatric facilities
Policies need to strengthen the implementation of patients’ rights, patient-centeredness, and transparency
Introduction
There has been limited systematic research to describe the quality of inpatient psychiatric care. 1 Most research on the quality of inpatient psychiatry has been conducted within single sites and focused on the use of restraint and seclusion.2,3 While restraint and seclusion are widely considered signals of treatment failure, serving as critical sources of both psychological and physical harm (including death),3‐10 they are not the only safety events to consider.
Complaints from healthcare providers, patients, and families have long been considered to be important barometers of safety, as well as sources of information about the diversity and nature of critical incidents (eg their types and antecedents). 11 Complaints have been used in quality feedback systems, 12 and in research to understand variation in care quality.13,14 Most of this work has been concentrated in general hospital or nursing home settings.13,15,16 There have been only a few studies to describe complaints made about inpatient psychiatric facilities, with these coming from outside the United States and from more than 10 years ago.17‐19
Institutions for Mental Diseases
In addition to understanding the frequency and types of complaints, it is of particular policy relevance to understand how safety indicators might differ between psychiatric facilities that are considered “Institutions for Mental Diseases” (IMD) compared to non-IMDs. IMDs are freestanding psychiatric facilities with more than 16 beds that serve adults ages 21–64. 20 IMDs are distinct from non-IMD facilities in that 1) they explicitly serve patients with psychiatric needs, while non-IMD facilities may be primarily focused on physical health, and 2) they are larger than non-IMD psychiatric-focused facilities, which may have a smaller number of beds. Since the inception of Medicaid in 1965, states have been prohibited from using Medicaid matching funds to pay for care in these facilities (known as the “IMD exclusion”). 20 The IMD exclusion has narrowed over time, with many states taking advantage of the Centers for Medicare and Medicaid Services’ (CMS) “in lieu of” guidance, 21 allowing Medicaid managed care organizations to receive matching funds for capitated payments if an enrollee is treated in an IMD setting for up to 15 days in a month. Calls to do away with the IMD exclusion altogether have implied that the IMD exclusion undermines healthcare parity and contributes to a shortage of psychiatric beds. 22 These advocacy efforts have led to the passage of the SUPPORT Act, which allows states to write Medicaid state plan amendments to reimburse IMDs for care, as well as the implementation of Section 1115 waivers that also permit Medicaid reimbursement of care in these settings.
However, this expansion of access to inpatient psychiatric care, through the narrowing of the IMD exclusion, is towards a very specific kind of facility. IMDs are large, freestanding psychiatric hospitals mostly owned by for-profit companies. 23 In contrast, most psychiatric facilities in the country are units of general hospitals, which are primarily nonprofit. 24 As such, IMDs may have different resources, missions, and business models than non-IMDs in ways that translate into differences in the quality of care provided. For example, both economic theory and empirical evidence from other healthcare settings suggest that for-profit healthcare providers are more likely to prioritize profits at the expense of quality.25,26 This behavior is likely to be exasperated during conditions in which patients lack the ability to make a choice or to “vote with their feet,” which is the case in inpatient psychiatry. 1 Prioritizing profit over quality might take the form of organizations employing inadequate staffing configurations, keeping beds filled at all times despite patients’ needs, and using behavioral containment methods inappropriately (eg, medications, restraint, and seclusion). Indeed, there is evidence that the largest providers of inpatient psychiatric beds in the country, primarily concentrated within IMDs, maintain considerable profits while employing these methods.27,28 These organizations stand to gain the most from the aforementioned policy changes to increase access to inpatient psychiatry, potentially incentivizing investment and growth in IMD beds. Indeed, evidence demonstrates a steady growth in for-profit ownership of psychiatric beds, concentrated among freestanding facilities, 24 with market analysts speculating that there will be continued investment by large corporations and private equity firms. 29
Current Study
The aim of the current study was to describe the types and frequency of regulatory-level safety indicators from inpatient psychiatric facilities in Massachusetts over time (2008–2018), and to identify differences in these indicators between IMDs and non-IMDs. We hypothesized that rates of complaints, restraint, and seclusion would be higher among IMDs versus non-IMDS. This is the first known study to describe variation in complaints from United States-based inpatient psychiatric facilities, as well as differences in complaints, restraint, and seclusion by IMD status.
Methods
Data Sources
Facility-level data on complaints, episodes of restraint and seclusion, discharges, and patient days for the years 2008–2018 came from the Massachusetts Department of Mental Health (DMH) through a series of public records requests. Information on facility characteristics was derived from the CMS Provider of Services (POS) files and checked against state licensing documents and online searches. This study took place from March 2020 to August 2021 and was approved by the Institutional Review Board.
Sample
The sample comprised all inpatient psychiatric facilities in the State of Massachusetts that were either licensed or owned by DMH from 2008 to 2018, including both psychiatric units of general hospitals and freestanding psychiatric hospitals (nearly all facilities in the state). We used a roster of licensed facilities by year, as well as annual counts of discharges and online searches, to determine whether a facility was open or closed in a given year.
Measures
Using data from the POS file, we classified facilities as IMDs if they served adults ages 18–64, were freestanding psychiatric facilities, and had more than 16 psychiatric beds. While the regulatory definition of an IMD is restricted to ages 21–64 (for Medicaid reimbursement purposes), hospitals are not organized around those ages but around standard age brackets for youth, adults, and older adults. All psychiatric facilities that serve ages 18–64 are, necessarily, IMDs because they include ages 21–64. Ownership was classified into three categories (nonprofit, for-profit, and government) using reported categories from the POS file.
Restraint and Seclusion
Episodes of restraint and seclusion came in files that reported the total number of restraints and number of seclusions, as well as rates of these counts adjusted for patient days. We report overall counts of these episodes and rates using 10,000 patient days as the denominator.
Complaints
Complaint data included counts of events that providers are mandated to report per licensing regulations, as well as other events from staff, patients, and families. These data came in a file with four descriptive elements for each complaint – the hospital associated with the complaint, the year of the complaint, a tag for the type of complaint, and the outcome of the investigation (substantiated, unsubstantiated). Tags for the type of complaint were preassigned by DMH, with a single complaint sometimes associated with multiple tags. We created two non-mutually exclusive buckets of these type tags; one focused on safety-related complaints (including negligence, sexual abuse, human rights violation, medicolegal death, staffing issues, physical abuse, restraint and seclusion, attempted suicide, medication issue, verbal abuse, and absent from the unit without authorization) and the other on abuse (including sexual, physical, and verbal abuse). Additional type tags included treatment issue, behavioral issue, missing/damaged property, environmental issue, policy issue, privacy issue/violation of the Health Insurance Portability and Accountability Act of 1996 (HIPAA), felony committed, and “other.”
We calculated summary rates of overall complaints, substantiated complaints, safety-related complaints, and abuse-related complaints. Rates were created by dividing each summary count by 10,000 patient days for a given facility in a given year, consistent with how we operationalized the restraint and seclusion measures. We did not restrict our analysis to substantiated complaints given that many factors go into substantiating a complaint, which do not necessarily indicate the validity of the complaint (eg, investigatory resources, ability to identify corroborating evidence). Prior work has used overall complaints (independent of the investigation's outcome) as signals for safety.13,30,31
Analysis
We report univariate statistics on the overall number of complaints and the composition of the different types of complaints, as well as total counts of restraint and seclusion episodes. We then report trends over time in the summary complaint rates (overall complaints, substantiated complaints, safety-related complaints, and abuse-related complaints) and rates of restraint and seclusion for a given facility, stratified by IMD status. Finally, we test for differences in means using a Welch's t-test and differences in medians using the Kruskal–Wallis equality-of-populations rank test. We test these differences with years pooled as there were no statistically significant changes in events across time.
Results
The majority of facilities were nonprofit (53.5%), followed by for-profit (35.8%), and government (10.8%). The overall median numbers of beds, discharges, and patient days per facility were 26, 777, and 8,836, respectively. IMDs were more likely to be for-profit and government-owned (61.4% and 27.4%, respectively) compared to non-IMDs (25.9% and 4.3%, respectively). IMDs had a much greater volume of patients per facility compared to non-IMDs, as indicated by differences in the median number of beds, discharges, and patient days (see supplemental material).
Across the 11 years, there were 92,670 restraint and seclusion events, with nearly 80% of these being restraint episodes (Table 1). There were 17,962 complaints, with the top ten related to treatment issues (count = 3,230, 18.0%), “other” (2,522, 14.1%), behavioral issue (2,422, 13.5%), human rights violation (2,073, 11.6%), physical abuse (2,047, 11.4%), verbal abuse (1,126, 6.3%), negligence (1,127, 6.3%), sexual abuse (945, 5.2%), medication issue (922, 5.1%), and missing/damaged property (613, 3.4%). Among the different types of complaints, 48.9% related to safety and 19.9% related to abuse.
Composition of Complaints and Their Distribution Across IMDs and Non-IMDs, 2008–2018.
Notes: Authors’ analysis of data from the Massachusetts Department of Mental Health in files received through public records requests, 2008–2018. The denominator in the “Percent of total” column corresponds with the value in each respective row under the “Total Hospitals” column. The summation of the different types is more than the overall number of complaints because a single complaint can have several “types” tagged. Sub is Substantiated. AWA is absent without authorization.
The distribution of episodes of restraint, seclusion, and complaints differed markedly between IMDs and non-IMDs. While IMDs comprised 57.3% of patient days, they represented 91.6% of complaints and 65.8% of restraint and seclusion episodes. When restricted to only private facilities, IMDs accounted for 45.9% of patient days and were responsible for 58% of complaints and 56.6% of restraint and seclusion episodes (see supplemental material).
In contrast to the overall composition of complaints, complaints about non-IMDs were primarily related to medicolegal death (266, 17.7%), negligence (230, 15.3%), human rights violations (220, 14.7%), sexual abuse (160, 10.7%), and treatment issues (163, 10.9%). Though IMDs had higher numbers of all types of complaints compared to non-IMDs, the composition of complaints about non-IMDs was more likely to skew towards being related to safety and abuse (73.5% and 22.7%, respectively) compared to the composition about IMDs (53.2% and 19.6%). Notably, these differences were driven by government-owned facilities. When restricted to private facilities, safety-related complaints comprised the vast majority of complaints about both IMDs and non-IMDs (88.8% and 90.2%, respectively), with IMDs having a greater share of complaints related to abuse (33.8%) compared to non-IMDs (26.3%).
There were no statistically significant changes in the events across time or between the years 2008 and 2018, though there was a non-significant increasing trend in episodes of restraint (see supplemental material).
Compared to non-IMD facilities, IMD hospitals had a 47.8% greater mean rate of restraint, 68.3% greater rate of seclusion, 276.9% greater rate of overall complaints, 284.8% greater rate of substantiated complaints, 183.6% greater rate of safety-related complaints, and a 236.1% greater rate of abuse-related complaints (see Figures 1 and 2 and supplemental material). When restricted to private facilities, the disparity persisted (see supplemental material).
Differences in mean rates of complaints for a given facility in a given year, stratified by IMD status. Notes: Authors’ analysis of data from the Massachusetts Department of Mental Health in files received through public records requests, 2008–2018. Statistical tests of difference were conducted between IMDs and non-IMDs using the Welch's Test for Unequal Variances of means. ***p < 0.001.
Differences in mean rates of restraint and seclusion for a given facility in a given year, stratified by IMD status. Notes: Authors’ analysis of data from the Massachusetts Department of Mental Health in files received through public records requests, 2008–2018. Statistical tests of difference were conducted between IMDs and non-IMDs using the Welch's Test for Unequal Variances of means. ***p < 0.001.
In Table 2, we calculated the average number of events for every 30 census days, and the average number of census days until an outcome occurs (see supplemental material). For every 30 census days in a given facility, a complaint was filed on average 0.9 times, restraint was used 7.5 times, and seclusion was used 1.8 times. However, for IMDs, a complaint was filed 4.4 times, restraint was used 21.3 times, and seclusion was used 5.6 times. For non-IMDs, a complaint was filed 0.3 times, restraint was used 4.1 times, and seclusion was used 1.0 times. The disparity between IMDs and non-IMDs was reduced when excluding government facilities, but remained large.
Average Number of Events per 30 Census Days and the Average Number of Census Days Until an Event Occurs, 2008–2018.
Notes: Authors’ analysis of data from the Massachusetts Department of Mental Health in files received through public records requests, 2008–2018. For details on how these values were calculated, see supplemental material 30
Discussion
Among Massachusetts inpatient psychiatric facilities, episodes of restraint, seclusion, and complaints were frequent occurrences between 2008 and 2018. Most of these events occurred in IMDs; these facilities had significantly higher event rates even after adjusting for patient days. While government facilities accounted for some of the large margins between IMDs and non-IMDs, a sizable disparity remained across all events when excluding government facilities. There were no differences in events across the 11 years, although there was a non-significant trend towards increasing restraint rates among both IMDs and non-IMDs. To our knowledge, this is the first systematic examination of complaints from multiple psychiatric facilities in the United States, and the first to examine differences in safety indicators between IMDs and non-IMDs.
The types and frequency of complaints observed in this study are concerning. In Massachusetts, nearly 20% of complaints were related to abuse, with about 2000 complaints related to physical abuse, 1126 related to verbal abuse, and almost 1000 related to sexual abuse across the 11 years. Research from other general hospital settings has reported an aggregate rate of 2.5% of abuse-related complaints, 32 far lower than what was observed in our study. Moreover, complaints may be undercounting the prevalence of these events. In a random sample survey of nursing home residents, it was estimated that nearly 25% experienced physical abuse from staff, 33 and a separate study that included resident and staff interviews estimated that about 20% experienced abuse from other residents in the span of just one month. 34 It is unclear how the nature and frequency of abuse differ between psychiatric patients and nursing home residents. In a study of within-hospital incident reporting systems at a sample of Veterans Health Administration psychiatric units, researchers estimated that only about a third of critical incidents identified in medical records were reported to the formal critical incident monitoring system. 35 The gap between true incidence of events and complaint indicators is likely even greater in our study, which is focused on those complaints that reach the state's regulator outside of the hospital.
In the state of Massachusetts, IMDs were much more likely to be for-profit or government-owned. Research on nursing homes has consistently found that for-profits perform worse on indicators of safety compared to nonprofits, including on measures of complaints.13,16,36 Given the strong association between ownership and IMD status, future research will need to tease apart the distinct contribution of ownership to quality of care. This is particularly relevant considering that for-profit psychiatric facilities have been increasing and might be most responsive to changes in IMD reimbursement policies.1,24 Indeed, a recent audit study of residential substance use treatment facilities found evidence of for-profits providing inferior quality of care. 37
Future research is also needed on IMD settings that primarily treat substance use disorders (SUDs), which differ from inpatient psychiatric facilities. Many states have had waivers to pay for care in IMDs for SUD for several years; thus, examining quality in these locations remains critical. One recent study identified that fewer than one-third of residential substance use programs for opioid use disorder offered opioid agonist maintenance therapy, 38 which is the standard of care for OUD. Whether or not there is a difference in IMD status remains to be examined.
We are unaware of any other research to examine quality variation across IMD status; ie, comparing large, freestanding psychiatric facilities to other psychiatric hospital settings. The results of our study call into question potential relative deficiencies in patient safety in IMDs, and challenge the current progression of advocacy and federal policies that have made it easier to use federal matching funds for care in IMDs. Policymakers should recognize the balance of harms and benefits that may result from these policy changes and commit to measuring and reporting these events to stakeholders.
To access the data used in this study, we submitted several public records requests and manually matched facilities to additional data. The data lack details about the events and people involved, as acquiring full complaints was cost prohibitive (ie, >$500,000) and would have been delivered in voluminous pages, rather than usable data elements. Patients and their families, therefore, have no reasonable way to understand potential differences in safety indicators across facilities. Other states, such as Colorado, report their complaints and inspection reports online; Massachusetts and other states should consider doing the same.
Our analysis indicates that the number of episodes of restraint, seclusion, and complaints did not decrease across the 11 years, suggesting that these data are not being effectively used in oversight and prevention work. In 2020, however, Massachusetts implemented a new complaint oversight system called the Safety Learning System. This new system has a more granular coding hierarchy designed to inform interventions. It will be important to understand how Massachusetts uses this system to affect change moving forward.
There is also a role for the federal government to motivate improvement and systematization in our data infrastructures. Medicare's Nursing Home Compare is a reasonable model to follow: Medicare compiles and reports information on federal inspection reports, penalties, staffing, and key quality metrics from nursing homes to support accountability and shopping. While there is room to improve the program, it is far more informative than what currently exists for inpatient psychiatry. The CMS's Inpatient Psychiatric Facility Quality Reporting Program (IPFQR) includes minimal information on safety and virtually no information on structural characteristics, such as staffing. While the IPFQR program does have measures of restraint and seclusion, there are weaknesses in these measures, as others have described elsewhere,39,40 and safety events extend beyond restraint and seclusion. Further, while there have been efforts to systematically collate and monitor state-level complaints related to nursing homes through the National Ombudsman Reporting System, no such clearinghouse exists for inpatient psychiatry.
Both regulators and payers should work towards motivating and supporting inpatient psychiatric facilities in identifying and addressing the root causes underlying these safety events. For example, when it comes to physical and sexual abuse, both patients and staff can be perpetrators and targets, but patients’ ability to seek recourse for these offenses is compromised due to the discrediting of their perceptions, power imbalance, and normalized cultures of violence. 41 Fundamental culture changes that center patients’ rights and patient-centered, trauma-informed care principles, might prove successful in both preventing and mitigating the consequences of these events.42‐45 Patient-centered care practices are evidence-based methods for preventing and reducing violence, trauma, and the use of restraint and seclusion in inpatient psychiatry.46‐48 There is some movement towards patient-centered care at the federal level. In 2022, congress mandated that CMS begin systematically collecting patient experience information from all inpatient psychiatric facilities participating in the Medicare program. In 2023, CMS responded with proposed regulatory rules requiring every inpatient psychiatric facility to administer a patient experience survey and report responses to CMS for public reporting.49,50 These policy changes can potentially move the field forward to better understand variations in care quality, the relationship between patient experience and safety events, and opportunities for intervention.
Limitations
We obtained only high-level information on counts of complaints and episodes of restraint and seclusion and were therefore unable to code for relevant information, such as who the complainant (eg, patient, staff, family, advocate) and antagonist/assailant was (eg, staff, patient), demographic and clinical information about patients, contextual information surrounding the nature of events, resulting injuries to patients and staff, and details on how issues were resolved (eg, reprimanding or firing staff). Relatedly, we did not have information on facility case-mix. However, we are not aware of any evidence to suggest that the types of complaints reported in this paper are justifiable among certain patient populations. Indeed, many of these complaints are “never events,” according to the Agency for Health Care Research and Quality, 51 which means they should not occur regardless of patient characteristics.
This study was restricted to one state. Massachusetts might differ in organization of services and complaint reporting systems; caution should therefore be made when generalizing these findings. Finally, the data reported in this study were not collected for research purposes and could be subject to error. The distribution of complaints should not be viewed as measures of prevalence of certain critical incidents; rather, they are likely to represent the “tip of the iceberg;” a diversity of methods is needed to describe the burden of different safety events in inpatient psychiatry.
Conclusions
Inpatient psychiatry has been excluded from the type of scrutiny that the rest of health care has experienced for decades now. 1 The results from this descriptive study emphasize the urgency with which we must improve the observability of inpatient psychiatry. The extent of harm being caused by these services that are intended to treat and support individuals experiencing psychological and emotional distress is a human rights concern that should be more seriously investigated by researchers, policymakers, regulators, and payors.
Stakeholders should be able to easily observe the rate and types of complaints and use of restraint and seclusion at facilities in their area without needing to submit public records requests. Researchers need access to improved data to empirically investigate market, systems, and organizational-level drivers of quality and to identify the best methods through which to motivate improvement. Policies should strengthen complaint reporting systems, transparency, patients’ rights, and patient-centeredness. These efforts must occur before or at least alongside policies that ease access to inpatient psychiatry, especially to IMDs.
Supplemental Material
sj-docx-1-jpx-10.1177_23743735231179072 - Supplemental material for Complaints, Restraint, and Seclusion in Massachusetts Inpatient Psychiatric Facilities, 2008–2018
Supplemental material, sj-docx-1-jpx-10.1177_23743735231179072 for Complaints, Restraint, and Seclusion in Massachusetts Inpatient Psychiatric Facilities, 2008–2018 by Morgan C. Shields and Mara A.G. Hollander in Journal of Patient Experience
Footnotes
Acknowledgements
The authors would like to thank Dr Meredith Rosenthal for her review of an earlier draft.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.
Supplemental Material
Supplemental material for this article is available online.
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
