Using the web-based Pennsylvania Patient Safety Reporting System (PA-PSRS), the PSA receives and analyzes reports of serious events – which involve clinical care and result in unanticipated death or injury of a patient requiring delivery of additional health care services, and incidents – which involve clinical care which could have injured the patient but did not cause unanticipated injury or require delivery of additional health care services. Serious events are simultaneously reported the to the state Department of Health (DOH), while a third type of report goes only to the DOH: infrastructure failures – which involve undesirable or unintended events involving a medical facility’s infrastructure, or the discontinuation or significant disruption of a service which could seriously compromise patient safety. The DOH retains regulatory and sanctioning authority as it has prior to PA-PSRS and can follow-up with facilities on an individual basis to ensure compliance with appropriate care policies.

The PSA has contracted with patient safety experts and clinical analysts – including pharmacists, nurses, equipment engineers and physicians – from ECRI and the Institute for Safe Medication Practices (ISMP) to comb through the data, prioritize and analyze it, follow-up with patient safety officers at individual institutions for more information as necessary, and issue Patient Safety Advisories, based on data gleaned from PA-PSRS and supplemented by medical and clinical literature, on steps to avoid future adverse events. The sixth quarterly issue of the Advisory is due this month. Individual facilities also have access to their own, real-time report trending data and statewide comparisons through PA-PSRS for their internal quality improvement activities.

Some interesting and potentially important details are presented in the PSA’s 2004 Annual Report, which also becomes the baseline with which future reports will be compared. Interpretation of the data, however, is not straightforward and may lead casual audiences to draw erroneous conclusions. An reported incident, for example, could be negatively interpreted as a medical error that almost happened, or positively interpreted as effective prevention. A serious event may involve a medical error, or may reflect an unanticipated adverse outcome not clearly attributable to error.

Critics argue that the PA-PSRS system has other significant shortcomings. PSA data are reported in aggregate form only and do not identify individual institutions, making comparisons between institutions impossible. PSA Advisory subjects are anecdotally presented and need a systematic research focus and peer-review credibility to spur greater buy-in. Although reporting is mandatory, enforcement is inadequate, and underreporting by facilities is likely. Baseline comparisons are unavailable to allow for meaningful interpretation of aggregate data. Facilities that report a high number of incidents and serious events are not held publicly accountable.

PSA Administrator Alan B.K. Rabinowitz believes most of these criticisms are misdirected, and notes that the PSA is meant to be an independent, educational, facility-based entity designed to retain health care provider confidentiality, offer expert advice on patient safety enhancement and encourage facilities to cultivate a culture of safety – in a non-regulatory and non-punitive fashion.

No information about individual patients and providers is collected through PA-PSRS, and no information about individual facilities is made public. According to PA-PSRS Clinical Director John R. Clarke, M.D., the PSA "has the responsibility of asking how the health care system can keep an unsafe act from harming a patient – no matter whether it is an honest, occasional mistake by an excellent provider or aberrant behavior by an unsafe provider."

PSA officials argue that the agency’s activities stem from a different paradigm from other initiatives that relate to patient safety, such as reports issued by the Pennsylvania Health Care Cost Containment Council (PHC4), regulatory and sanctioning activities by the DOH and medical boards, best practice registries cultivated by the Pittsburgh Regional Healthcare Initiative, and malpractice lawsuits filed within the legal system. The PHC4 serves a public accountability function by issuing report cards of individual health care institutions using outcome benchmarks; regulators serve an enforcement function by policing inappropriate provider behavior; and lawsuits seek restitution, punishment, and perhaps deterrence against medical malpractice.

The PSA says its activities and goals are fundamentally different from those other efforts, while working in tandem with them. For example, the PSA believes that benchmarking performance of individual hospitals relative to other hospitals is far less important than spurring process improvement among hospitals, irrespective of event frequency. The degree to which hospitals are using PSA feedback varies: all receive the PSA advisories, some are actively using the PA-PSRS trending data tool, while others still prefer their own internal risk management and quality improvement tools.

New Data

The 2004 Annual Report noted that 427 facilities – including 249 hospitals, 173 ambulatory surgical facilities and five birthing centers – were required to submit data to the PSA, and that 98.7 percent of all reports came from hospitals.

PA-PSRS classifies nine basic event types – medication errors; adverse drug reactions (not a medication error); equipment, supplies or devices; falls; errors related to procedures, treatments or tests; complications of procedures, treatments or tests; transfusions; skin integrity; and other/miscellaneous – each of which is broken down into second- and third-level subcategories, while the complete taxonomy has 198 distinct event types. Data are reported as statewide aggregate figures, and are broken out by six regions, although in limited fashion, in the Annual Report.

Highlights from the fist six months of reporting include the following:

· 70,851 reports were logged between June 7 and Dec. 31, 2004, of which 67,104 (95 percent) were incidents and 3,747 (five percent) were serious events.

· Of the serious events reported, 207 involved a patient death (0.3 percent of all reports and 5.5 percent of all serious events).

· The most frequently reported type of serious event was complications related to procedures, treatments or tests – accounting for 31 percent of serious events, and 58 percent of all deaths.

· Medication errors accounted for 25 percent of all reports (the most frequently reported category), one percent of all serious events, and 2.9 percent of deaths.

· Falls accounted for 21 percent of all reports (the second most frequently reported category), and 22 percent of all serious events (second most frequently reported serious event).

· Patients over the age of 65 were involved in 51.2 percent of all reports and 59 percent of all serious events reported.

The report further details findings within event types. For example, the most frequently reported serious event related to surgical errors (53.8 percent) was "unintended laceration or puncture," which most frequently occurred during colonoscopies and hysterectomies. The PSA said it plans to address the issue in an upcoming advisory.

The pace of reporting appears to be accelerating, particularly reporting of serious events. As of early May, the PSA has received about 130,000 incident and serious event reports, according to Rabinowitz. Between July 1, 2004 and April 30, 2005, PA-PSRS had collected 20,601 serious event reports, which prompted the DOH to initiate 457 telephone follow-ups with institutions and 383 on-site visits, according to Richard Lee, DOH’s deputy secretary of quality assurance. Citing data from a different time frame – between July 1, 2004 and Feb. 28, 2005 – Lee notes that site visits prompted by PA-PSRS reports resulted in the DOH identifying 12 deficiencies, each requiring facilities to submit a plan of correction to the department.

Lee notes that the numbers are not comparable to the DOH’s prior Chapter 51 reporting requirements, which he says used a less precise definition of events reportable to DOH, and which was fraught with reporting difficulties and inaccuracies. As per Act 13, PA-PSRS has replaced Chapter 51 hospital reporting of serious events, and the DOH is getting more complete reporting and precise information than before, he notes.

Lee also notes that, between July 1, 2004 and Feb. 28, 2005, the DOH received 11,803 reports of infrastructure failure.

The PSA report urges "considerable caution" when interpreting PA-PSRS data and warns that raw numbers by themselves are meaningless without knowing the number of patients seen, the inherent risks of the procedures undertaken and the diligence of the facility in finding and sharing information about unsafe actions and bad outcomes – all of which are unknown. While volume of reports by any one facility or group of facilities is not made public, such information would be misleading in any case, as the data are not adjusted for these factors.

The report maintains that PA-PSRS data are inappropriate for use as a "report card" for other fundamental reasons. A facility with substantially more reports than its peers could be regarded as safer because it may have better systems in place for recognizing and reporting actual and potential adverse events. The number of incorrect medications reported is also not meaningful without knowing its ratio to all medications administered. Further, incidents and serious events do not necessarily involve error or unsafe action. Administering the wrong medication is an error, while an adverse reaction to a drug the patient never received before is not a error, although it is unanticipated and therefore reportable. An emergency operation on a patient with a ruptured aortic aneurysm who dies may be reported as a serious event, even if the procedure is associated with a 50 percent mortality rate.

The most meaningful statistic, according to Clarke, is the "recovery rate" – the percentage of reported events that are not associated with harm, i.e., a low number of serious events, a high number of incidents, or some combination. Reported event types that had a high ratio of serious events to incidents (significantly higher than the overall average of five percent) include adverse drug reactions, complications of procedure/test/treatment, and skin integrity. Reported event types that had a low ratio of serious events to incidents include medication errors, equipment/supplies/devices, and errors related to procedure/test/treatment.

The report warns that valid comparisons between PSA data and patient safety data from other sources are difficult because of Act 13’s unique definitions of what is and what is not reportable to PA-PSRS, while the system’s event types are also specific to PA-PSRS – including the nation’s only mandatory reporting of "near misses" which did not harm patients. The report notes that the National Quality Forum is currently developing a national taxonomy for patient safety reporting that may enable PA-PSRS to match Pa.’s data with data from similar reporting systems.

The report acknowledges other definition-related wrinkles in the data. Many facilities have expressed to the PSA their difficulty in determining whether complications suffered by a patient were "unanticipated" – the threshold for what is reportable, while facilities are excessively using the "Other" event type category when they have difficulty classifying a report by another event type. PSA says it is working to clarify these issues.

Critical Observations

Enthusiasm over how effectively the PSA is enhancing patient safety varies across stakeholders, ranging from qualified praise to outright skepticism.

From a consumer standpoint, the PSA represents one of the better efforts in the country to detect errors and improve patient safety, particularly with its mandatory "near-miss" reporting, which has been shown to be an effective safety measure in industry, according to Barry Boughton, from Public Citizen’s Congress Watch Division. He believes the 70,000 incidents and events tallied in the PSA’s 2004 Annual Report indicates a fair level of reporting compliance and, of the 22 or 23 states that have some form of mandatory patient safety reporting, he doesn’t remember seeing any with that many reports.

Boughton also likes PSA’s separate funding mechanism, which is drawn from an assessment on health care institutions, rather than depending on the less-secure whims of annual appropriations by a Legislature.

Boughton is critical of the PSA’s "overblown concern" with confidentiality, especially for serious events, which he believes merit disclosing the names of institutions and doctors to allow public comparison across institutions, and to spur greater accountability and deter errors within institutions. He is skeptical that confidentiality is necessary to create a "culture of learning," and he points to literature that disclosure and apologies to patients for mistakes do not increase the risk or cost of litigation to an institution.

From a legislator’s standpoint, the 70,000 reports show what has been suspected all along: that medical errors can be problematic for health care facilities and patients, according to Tom Andrews, press secretary for Minority House Leader H. William DeWeese (D-Fayette/Greene/Washington). Without Act 13, we would have never had these data, and it is now up to the institutions to address the errors in order to sustain their reputation as providers of safe and high-quality health care, he adds.

From a physician perspective, the PSA’s sizeable number of reports and its informative advisories indicate that the agency is accomplishing its goals well, according to Jeff Greenawalt, director of Public Health and Professional Licensure for the Pennsylvania Medical Society (PMS). Perhaps the most significant evidence of impact, he says, is that 30 percent of hospitals surveyed by the PSA report having made some change as a consequence of the advisories. Greenawalt says he has not heard of any negative impact on, or complaints by physicians regarding PA-PSRS.

While the PMS was not an advocate for creation of the PSA, says Greenawalt, the PMS wanted to see it function as safety improvement systems in the aviation industry function – as nonpunitive, confidential learning tools. The system would fail and reporting would dry up if reported errors were exposed to the threat of sanctions and license revocations, he says, noting that he has yet to hear of a case identified through the PSA in which an individual physician was sanctioned or subjected to disciplinary action. Act 13’s requirement that institutions report serious events, in writing, to patients and/or their families within seven days of discovery has not generated concern among physicians, and was already standard practice by hospitals, at least verbally, says Greenawalt.

Confidentiality does not sacrifice accountability, Greenawalt believes, because most patient safety problems are not competency issues, but are system issues that can be prevented in the future, not by sanctioning an individual provider, but by flagging all providers to be cautious about the issue. For that reason, Greenawalt says, the disciplinary functions of the DOH and medical boards must be kept separate from the educational function of the PSA.

Clifford Rieders, Esq., who was president of the Pennsylvania Trial Lawyers Association at the time Act 13 was passed, says he had insisted that Act 13 include a patient safety component before PaTLA would agree to the Act’s tort reform provisions. Rieders says he helped draft the legislation, and as a PSA board member has been an active member of all its subcommittees. Although a knowledge of the incidence of medical errors is crucial to improve patient safety and decrease malpractice lawsuits, Rieders says, "nobody will claim that the PSA has identified and improved very much in those first six months. We are in kindergarten, in terms of patient safety in Pa." The system shows promise, but needs further development to meet its goals and overcome hospital resistance, he argues.

While 70,000 reports is nothing to be scoffed at, Rieders maintains that hospitals are under-reporting, and he points to a significantly higher number of reports – and a greater proportion of incidents to serious events – from northcentral Pa. hospitals than from hospitals in other regions of the state, when adjusted for volume of health care provided (reports per 1,000 patient days). Such a discrepancy, he says, is likely evidence of under-reporting elsewhere.

Rieders finds it odd that, despite the reliance on whistleblower protections for persons who anonymously inform the PSA of a facility’s failure to comply with mandatory reporting requirements, only one valid anonymous report of noncompliance was received during the six-month interval of the PSA’s report. That institution was subsequently allowed to supply the information without sanction.

There is also no way to track the degree to which hospitals are complying with their mandate to report in writing serious events to families within seven days of discovery, and Rieders says his law firm, which reviews about 500 medical malpractice cases per year, last year saw only one case in which such a serious event notification was reviewed.

Rieders further believes that the public has a right to know details about reports from specific hospitals, which he says would further spur self-improvement through accountability and competition, and outweighs the risk of squelching hospitals’ reporting.

A more serious deficiency in the system is that the PSA does not have the staff to investigate every death, and its investigators have spot checked only 60 out of the 207 reported deaths, says Rieders.

"If you have honest reporting, aggressive enforcement and more complete analysis, that would encourage reporting by successful performers," says Rieders.

From a purchaser perspective, the PSA’s ability to share root-cause information with hospitals is an important part of the process of patient safety but, "as a stand-alone agency, this safe place to share information also becomes a safe place to do nothing," says Cliff Shannon, president of SMC Business Councils, which represents 4,000 employers, primarily in southwestern Pa. "I am indicting the idea that the PSA is going to catalyze and facilitate significant improvement in patient safety," he says, adding: "My skepticism is that the creation of the PSA was politically calculated to let off steam from the medical malpractice debate."

Shannon says the PSA lacks any mechanism that purchasers can rely on to ensure that hospitals will use data to improve patient safety in any measurable way. "When health care premiums are approaching $2,000 a month and families are still imperiled, ‘Take our word for it’ is not enough. I do not trust the ‘power’ of retrospective reporting of numbers of events, especially when they are not validated, as proof of improvement."

Shannon is further skeptical that mandatory reporting systems provide sufficient compliance incentive, and he points to a recent law that requires reporting of hospital-acquired infections to the PHC4, and which has produced 12,000 such reports for all of 2004 – a figure he says is at least ten times lower than billing records and national literature say it should be. "The PSA needs to be appreciated against the backdrop of failure to report by most hospitals in the state, despite a legal requirement that they report accurately," says Shannon. "The PSA has no such teeth. What am I, as a purchaser, to assume about methodical, effective improvement by the PSA?" he asks.

From a quality improvement standpoint, the PSA has "done what they’ve said they would do in its first year of operations: establish a system to get hospitals to report, and raise the public’s consciousness of patient safety," although significant challenges remain, according to David B. Nash, M.D., MBA, chairman of Jefferson Medical College’s Department of Health Policy. Nash believes that the event taxonomy and definitions being used by the PSA seriously limit its ability to compare data with other states, and he points to the difficulty of hospitals understanding what "unanticipated injury" means, as the threshold of what must be reported. The PSA should instead embrace a better classification system for medical error reporting, which is being developed at the national level by the National Quality Forum – involving hundreds of experts around the country – and is scheduled to be released this June, says Nash. A standardized taxonomy with national stature is critical, he believes, if Pa. hopes to tap into patient safety initiatives elsewhere, and to allow for state-to-state comparisons and national benchmarking.

A longer-range challenge for every state-based patient safety system is to "close the feedback loop to providers" and establish a formalized way to evaluate the impact of its advice, says Nash. "How do we know the PSA will improve safety? They need better tracking of the impact of their advisories, and other strategies to confirm that strategies for safety improvement are being used," he says.

Nash believes that the PSA needs a research agenda to improve the precision of its efforts and magnify its influence, including peer-reviewed publications to "bring traction" and clout to its advice, which he says has so far been anecdotal in nature. There is no way of knowing how many errors it takes, or what criteria are used to trigger an advisory, says Nash.

While the PSA and DOH rely on whistleblower protections for those who identify violations in facilities’ reporting obligations, and the DOH may levy sanctions against violators, Nash says there is no way to know whether hospitals are completely and accurately reporting to PSA. That only 2.9 percent of PA-PSRS-reported deaths were attributed to medication errors, compared to published national estimates of seven percent or higher, "speaks to the probability of some underreporting," Nash believes.

While Nash says that a whistleblower-based compliance assurance mechanism is the "state-of-the-art," with no superior system having yet been adopted elsewhere, it is insufficient and its shortcomings should be recognized.

PSA’s Appraisal

Given that the PSA’s primary focus thus far has been to get up and running, which entailed a cumbersome and bureaucratic bidding process to select vendors with whom to contract, PSA Administrator Rabinowitz is pleased with what it has accomplished so far. "We didn’t know what volume of reporting to expect, as there is no precedent in other states," he says, "but we think the volume we’re getting is a significant amount of reporting. We’re not focusing on what is not reported. What we are receiving is an acceptable volume of reports for a new system, and we hope to see an increase in reporting," Rabinowitz says.

Each facility is responsible for determining what meets the threshold of reportable incidents and events and, although the definition of "unanticipated injury" remains a source of confusion and inconsistency for some institutions, each must establish its own protocol to defend its reporting behavior, says Rabinowitz. The PSA is working with patient safety officers of institutions to clarify interpretations of the definition and, with only ten months of data collected so far, Rabinowitz says it would be premature for the PSA to conclude that the definition is interfering with complete and accurate reporting. He notes that the PSA could recommend that the Legislature reappraise the definition in the future, if necessary.

Rabinowitz believes that institutions have incentive to want to report incidents and serious events: "The more information facilities submit, the more information they have for their own quality improvement efforts. Patient safety will be enhanced when institutions recognize the value of the data." He reiterates that the PSA as a learning organization does not play a role in ensuring reporting compliance, and that it is the DOH’s responsibility to sanction facilities for noncompliance.

The DOH relies on two mechanisms to ensure reporting compliance, according to Lee: complaints of non-reporting, e.g., from a patient, family member or nurse; and the DOH’s scheduled site visits, during which inspectors pull a random sample of patient charts and, if they find incidents or events that were unreported though PA-PSRS, the DOH can issue sanctions. To the extent that the DOH discovers a problem involving inappropriate behavior of a physician or nurse, it will pass on the information to the state licensing boards. Lee says he is not aware of these things happening since PA-PSRS has been up and running.

Further buy-in incentive for hospitals may accrue from a provision in Act 13 that provides for a discount on an institution’s medical malpractice premiums if it can demonstrate a reduction in serious events as a result of adopting a program recommended by the PSA, which has since identified the Stand Up for Patient Safety Program, and the 100,000 Lives Campaign as two such choices. Details of the discount offer remain to be worked out by the Pa. Insurance Department and the DOH.

Rabinowitz believes that PA-PSRS has gained significant credibility and buy-in among institutions because of the respect and attention the PSA has paid to their concerns while the system was being developed and piloted – including offering 19 training sessions at 11 locations around the state, and a commitment to develop an interface between PA-PSRS and facilities’ existing reporting systems to minimize redundancy in data entry, says Rabinowitz.

Still, the PSA’s board is examining its goals and activities for the next few years, including ways to encourage full and open disclosure by facilities, to promote research, and to collaborate with other organizations around the state. The board is looking at ways to expand its research and scholarship, including the possibility of a peer-review mechanism for its advisories, making its data available to researchers, and running educational outreach symposia and CME sessions sponsored by the PMS and by individual hospitals, says Rabinowitz.

The methods used by PSA’s team of full-time analytical staff to triage data for follow-up are working well, says PA-PSRS Clinical Director Clarke. The team uses a cost-benefit analysis to determine which data are important to investigate further and what subjects merit inclusion in advisories. Analysis criteria include frequency and severity of an event, how representative the event is of a broader issue, how likely the team can develop a solution for it, how feasible it would be for institutions to implement the solution, and what kind of improvement can be anticipated if the solution were implemented, according to Clarke. The team meets weekly to identify "high-impact areas that would make a difference," and a diversity of content is selected for publication in the advisories to be useful to a diverse audience of facilities, he adds.

Sometimes only one or two reports could trigger a follow-up, as was the case when analysts discovered that identical packaging was being used for tuberculin and insulin syringes, which could result in an insulin overdose if the wrong syringe were used. Analysts thought the issue important enough for the PSA to issue its first supplemental advisory, rather than wait for the information to appear in the next quarterly issue, says Rabinowitz. One institution reported a syringe-related near-miss within one hour of receiving the advisory via email, and the PSA has since learned that several dozen institutions encountered the syringe packaging confusion, he adds.

PSA analysts tap several sources to generate solutions to problems, Rabinowitz notes, including reports from facilities themselves indicating what they did to avert patient harm, libraries and databases garnered by ECRI and ISMP during their national patient safety operations, and literature reviews which appear as footnotes in the advisories. Rabinowitz notes that PSA solutions are mostly gleaned from the "near-miss" incident reports which, as "serious events waiting to happen," are regarded by patient safety literature to be the key source of opportunity to learn how events are averted. A fair amount of solutions the PSA has offered so far are essentially reminders of safe practices that are widely-know but not followed, such as measures to avoid patient misidentification, he adds.

Clarke acknowledges that analysts do not yet have a way of comparing Pa. data to other states’ or national data, as PSA analysts could not find a statistically valid benchmark for comparison. Because PA-PSRS’ legal definitions of serious events and near-misses are essentially proprietary, "we’re starting out comparing ourselves to ourselves," he says. While the PSA is trying to find ways to correlate its event categories to those under development by the National Quality Forum (NQF), the effort is difficult because of fundamental differences between the two taxonomies. Whereas PA-PSRS reporting categories are based on initial presentation of problems – such as medication error types; locations where falls occur; and specific problems with clinical procedures – the NQF’s intentionally broad categories are designed to cluster different types of problems into underlying causes – such as communication problems, which could cover several sorts of clinical procedures, Clarke explains.

While the PSA had anticipated that a national standard for patient safety reporting would eventually emerge, and wanted to have interoperability with such a standard during its vendor RFP two years ago, NQF’s taxonomy had not yet been sufficiently developed and the PSA had to be responsive to Act 13 requirements without waiting for a consensus taxonomy to emerge, says Rabinowitz.

The taxonomy chosen by the PSA is proving useful and, to the extent that value can be added by mapping to NQF’s categories, the PSA will do so, adds Clarke. He candidly says, however, that making national benchmark comparisons is not only of secondary importance to PSA’s goal of spurring patient safety improvement among Pa.’s health care institutions, but may also be misleading and inappropriate. Real improvement, he argues, does not come from comparing performance to averages – national or otherwise – but comes instead by adopting the process-based paradigm of pursuing optimality, for example, eliminating all wrong-site surgeries or all nosocomial infections. "The best performance one can expect under optimal circumstances: that’s the benchmark," he says.

Impact on Hospitals

The ultimate measure of the PSA’s success will be the degree to which health care facilities actually use its data to improve patient safety. That appears to be happening, at least to some extent, among Pa.’s hospitals. More than 30 percent of all hospitals responding to a PSA survey (with a 62 percent response rate) indicated that they have implemented patient safety protocols as a result of specific articles in the advisories. The PSA’s 2004 Annual Report also offers anecdotes of specific changes hospitals have made as a result of information they used from the advisories, such as: training to avoid fires attributed to oxygen pooling in the operating room, improving MRI screening to include implantable devices, converting to latex-free products when such a choice is available, and tracking transdermal patches to prevent overdosing.

The Hospital & Healthsystem Association of Pennsylvania (HAP) says it is very pleased with how the PSA is functioning and says it receives positive feedback about the advisories from hospital patient safety officers, according to Paula Bussard, vice president of policy and regulatory services. HAP has not conducted a survey of its own to track whether hospitals have made specific changes as a result of the advisories, or whether hospitals are using the PA-PSRS trending data tool, but may learn more about those activities at a patient safety forum it is holding this month, she says. Bussard believes the trending tools would be particularly useful to medium and smaller hospitals that did not have their own patient safety trending systems before.

PA-PSRS is changing behaviors at Pa.’s hospitals in various ways, and to various degrees.

Lebanon’s Good Samaritan Hospital was among 10 hospitals in Pa., and 135 hospitals nationwide, that recently received a Distinguished Hospital Award for Patient Safety from HealthGrades, and attributed its performance to increased awareness of patient safety issues, for which it explicitly credited the PSA, according to the Harrisburg Patriot-News.

Jersey Shore Hospital in Lycoming Co. uses PA-PSRS data at its monthly patient safety meetings, reviewing every event and conducting root cause analyses if it sees something that poses risk to patients, according to the hospital’s patient safety officer, Betsy Crossley. Based on one such review, the critical access hospital streamlined its patient transfer process, which previously required contacting individual specialists for each patient transfer, by identifying a hospitalist at Susquehanna Williamsport Hospital to facilitate all transfers, she says, also noting that PSA’s reporting requirement has motivated the hospital to give its staff better guidelines on how to report events, to whom, and to stress the process of timely review.

Somerset Hospital says that the web-based PA-PSRS trending tool offers useful ways to track incidents and events, such as by unit or time period, and gives the hospital a mechanism of comparing its performance against its peers – which helps affirm the hospital’s ongoing root cause activities, says Craig Saylor, assistant vice president of quality and patient safety officer. The PA-PSRS trending tool is much more convenient than the paper-based reporting system the hospital previously used, he notes. Staff awareness of PSA advisories also gives tremendous clout, he adds, to safety improvement measures that the hospital had in place before PSA feedback – such as using preprogrammed IV pumps, automated dispensing equipment, a falls prevention program, and a hospital-wide handwashing program.

Based on communications with colleagues elsewhere, Saylor believes that hospitals may be over-reporting incidents and events through PA-PSRS, for fear of being audited and sanctioned, while a "blame-free" culture change may also be encouraging staff members to discover and report more than they have in the past. Somerset’s volume of events reported has increased because of education since the system’s implementation – averaging 14 incidents and events per month in FY05, compared to nine per month in FY04, says Saylor.

Diane Christie, patient safety officer at Riddle Memorial Hospital in Delaware Co., says she copies articles from PSA advisories and distributes them to different department directors, or in some cases, chiefs of surgery. Although Christie says the advisories have so far generally covered areas that Riddle had already addressed, they help hospital staff become more aware of their procedures and tweak them if necessary, for example, putting medical staff and their patients on alert to the possibility of colonoscopy injuries.

Riddle is also using the PA-PSRS trending tool to study its medication administration process on a monthly basis and has noticed that its medication errors are decreasing, Christie says, noting that the "Contributing factors to the event" section is particularly helpful in reviewing causes, and makes the new system superior to a less-detailed risk management tool the hospital had been using.

PA-PSRS reporting and the PSA advisories are reinforcing the importance of safety enhancements already in place at Mount Nitanny Medical Center, and are raising staff vigilance in identifying potential problems, according to Frank Speidel, M.D., the hospital’s senior vice president of medical affairs and chief quality officer, who notes that the hospital is seeing 40 percent more reports identified over the past two years.

Mount Nitanny is a member of CHART, a 33- hospital risk retention group malpractice insurance provider in central Pa., and is using CHART’s software instead of the PA-PSRS trending tool to conduct risk management benchmarking, says Herb Wilson, the hospital’s risk manager and patient safety officer. Wilson says he is more comfortable using CHART’s tool, which also allows benchmarking against a group of similar hospitals using CHART’s database. Mount Nitanny is double-reporting patient safety events, and Wilson says he will be glad when PA-PSRS is able to interface with CHART’s system to eliminate the reporting duplication burden.

University of Pittsburgh Medical Center circulates PSA advisories among patient safety officers at various institutions throughout the UPMC Health System and, although they have had minimal impact thus far, they can help a hospital make changes to avert possible problems, says Richard Kidwell, J.D., UPMC’s director of risk management. At least one or two UPMC hospitals made changes as a result of an item in the March 2005 Advisory about dangers associated with unlabeled basins, bowls and cups on the surgical field, he notes.

While UPMC hospitals collect and trend patient safety data using their own internal system, Kidwell says it is useful to get information from a wider spectrum of institutions, and he notes that some UPMC hospitals are using PA-PSRS data for such comparisons.

The PSA’s 2004 Annual Report is very helpful, as institutions generally don’t share medical error data across health systems, and the report reveals regional trends across the state, says Daniel Feinberg, M.D., patient safety officer and director of informatics at Pennsylvania Hospital. He says he has not used the PA-PSRS trending tool, since the University of Pennsylvania Health System has its own web-based incident reporting system that was developed by Penn and allows for sophisticated trending analyses. Penn plans to look at the PA-PSRS tool to see if it can add useful information to Penn’s internal system, such as seeing how they compare to statewide and regional data, then using its internal system to run reports based on trends identified by PA-PSRS and bringing the information to meetings of Penn Health System’s patient safety staff, he adds.