The Effect of CMS’ Inpatient Psychiatric Facility Quality Reporting Program on Nontargeted Indicators of Safety in Massachusetts

Research Objective In 2012, CMS implemented the Inpatient Psychiatric Facility Quality Reporting Program (IPFQR), which publicly reports facilities’ performance on a suite of measures, including targeted measures of restraint and seclusion (R/S). To perform, facilities could respond with approaches that are harder (eg, improving staff‐patient relationships) or easier (eg, increasing sedatives). If a facility responds meaningfully, this could have positive spillover on other aspects of quality outside R/S use. If a facility does not respond meaningfully or at all, the IPFQR program could negatively impact quality given the resources needed to comply with the reporting program (ie, burden on frontline staff). Given lack of choice and weak incentives to perform, economic theories suggest for‐profits would be more likely to exhibit a negative spillover. This study aims to evaluate the differential spillover effect of the IPFQR program on non‐targeted (but proximal) indicators of safety (ie, complaints) between for‐profit and nonprofit inpatient psychiatric facilities in Massachusetts. Study Design Episodes of R/S (targeted), complaints (nontargeted), and discharges (used to create rates) were obtained for 2008‐2017 through a series of public record requests made to the Massachusetts Department of Mental Health. Five summary scores were created using the complaint data: 1) overall complaints; 2) substantiated complaints; 3) safety‐related; 4) abuse‐related (verbal, sexual, physical); and 5) R/S‐related. Univariate and bivariate analyses were conducted using nonparametric methods given skewedness. A difference‐in‐differences estimator was employed using a Poisson distribution, with discharges as the exposure. A fixed effect for the facility was included to control for time‐invariant differences between facilities and errors were clustered on the facility. Coefficients are reported as incident rate ratios (IRRs). Population Studied For‐profit and nonprofit psychiatric facilities in Massachusetts who participated in the IPFQR program (N = 50 unique facilities, 490 observations for 2008‐2017). Principal Findings Results from the regression models demonstrate that for‐profits had a greater increase in overall complaints (IRR = 1.47, P  = .002), safety‐related complaints (IRR=1.51, P  = .005), abuse‐related complaints (IRR = 1.59, P  = .026), and, most strongly, R/S‐related complaints (IRR = 3.34, P  = .002). There was no statistically significant relative increase in substantiated complaints or episodes of R/S use. A visual inspection of complaints across years suggests clear parallel trends in the preperiod and clear divergence in 2013. Segmented regressions with placebo intervention dates support these findings. Conclusions There was a greater increase in complaints, especially those related to the target of R/S, among for‐profits relative to nonprofits, but no change in the targeted construct of R/S use. This could be due to the burden of complying with the IPFQR program on frontline staff and resource constraints. Ad hoc analyses of death data suggest this is a true change in quality rather than an increase in reporting. There are several limitations, including lack of information related to each complaint and lack of patient‐level data. Implications for Policy or Practice This is the first study to examine the effect of a national public reporting program on non‐targeted quality among inpatient psychiatric facilities, as well as variation across ownership. The findings point to the need to improve our data infrastructure to support tracking of intended and unintended consequences of policy interventions on quality of inpatient psychiatry. Primary Funding Source National Institutes of Health.