INVOLVING THE PUBLIC IN ASSESSING HEALTH QUALITY MEASURES Shoshanna Sofaer, Presenter Kirsten Firminger, Research Associate.

INVOLVING THE PUBLIC IN ASSESSING HEALTH QUALITY

MEASURES

Shoshanna Sofaer, Presenter

Kirsten Firminger, Research Associate

December 2 2009 School of Public Affairs Faculty Colloquium

2

PRESENTATION OVERVIEW

• Project • Research Team• Methods• Findings• Implications• The Sequel


3

THE PROJECT

• “Developing and testing nursing quality measures with consumers and patients”

• Funded by the Interdisciplinary Nursing Quality Research Initiative (INQRI) of the Robert Wood Johnson Foundation

• Project had three aims; we will focus on the first: Gauge public response to 15 nursing sensitive quality measures endorsed by the National Quality Forum (NQF)


4

The Interdisciplinary Research Team

• Shoshanna Sofaer, DrPH, Baruch College School of Public Affairs (Baruch), Principal Investigator

• Jean Johnson, PhD, RN, FAAN, George Washington University School of Medicine and Health Sciences (GW), Co-Principal Investigator

• Ellen Dawson, PhD, RN, ANP (GW), Co-Principal Investigator


5

The Interdisciplinary Research Team

• Kirsten Firminger, doctoral candidate, (Baruch)

• Christine Pintz, PhD, RNC(GW)• Andrea Brassard, DNSc, MPH, ANP(GW)• Kirsten and Christine actually ran the focus

groups we describe, coded the transcripts, and analyzed the quantitative and qualitative data


6

Background/Motivation

• Public reporting of quality measures is on the rise, but many public reports are not looked at, not understood and not used by the intended audience

• We know that unless quality measures resonate with the public, are easy to understand and apply, they will not influence behavior

• Yet many measures that meet technical criteria have not been tested with the public


7


• The National Quality Forum is a non-profit membership organization governed by multiple health care stakeholders

• They utilize a complex process to reach consensus across stakeholders on health care quality measures that meet certain criteria (next slide)

• When consensus is reached, the measures are “endorsed”• Many major sponsors of health care quality reporting will

only publicly report NQF endorsed measures, including the Centers for Medicare & Medicaid Services


8


• The criteria used by NQF are:– Importance to Measure and Report to make significant

gains in healthcare quality and improve health outcomes

– Scientific Acceptability of Measure Properties, i.e. validity and reliability

– Usability by intended audiences (e.g. consumers, purchasers, providers, policy makers) i.e. can understand and would find useful in decision making

– Feasibility of data collection without undue burden


9


• However, while extensive documentation is required about the scientific reliability and validity of measures, and feasibility of data collection, no documentation is required to demonstration the usability of measures to a major audience for public quality reports: consumers

• Instead, NQF relies on the views of members who are considered to be “representing” consumers, such as the AARP, the AFL-CIO, etc.


10


• This research was designed to investigate the hypothesis that at least some endorsed measures are, when actually tested with consumers, less than usable

• We hoped that if our hunch proved accurate, we could use our results to influence the process by which NQF endorses measures going forward

• The focus of our work was a set of 15 measures that were endorsed by NQF as being “nursing sensitive” hospital quality measures: the NQF 15


11

Research Methods

• We conducted 9 focus groups with recently hospitalized patients to determine if participants:– Could understand and interpret the NQF 15 – Found these measures significant in the light of

their hospital experiences– Thought the measures related to the roles,

responsibilities and functions of nurses in particular


12

Research Methods

• Sampling:

– Three groups each in DC, Chicago and Phoenix

– Three groups each with people who used OB, medical and surgical services

– Participants recruited to vary in age, gender, race/ethnicity, education level


13

Research Methods

• To maximize understanding of measures, we developed more user friendly language for the NQF 15 (see handout)

• To ensure comprehension of the task, we presented slides on quality measures

• To gauge understanding, we asked participants to explain the measure in their own words


14

Research Methods

• Participants silently rated each measure on their:– Importance – Likelihood of variation in performance across

nurses and hospitals – Whether a very high or low score would lead

them to reconsider their hospital choice• After each rating, participants discussed

their responses and reasons for their ratings in depth


15

Findings

• On silent ratings, at least 80% of participants found the following measures very important:– Failure to rescue– Pressure ulcers– UTI among catheterized patients– Central line infections– Pneumonia among patients on ventilators– Positive nursing work environment


16

Findings

• Ratings showed participants were least responsive to three measures about whether nurses advised patients to stop smoking

• Group discussion generated these reasons: – Nurses have better things to do with their time– Patients already know they shouldn’t smoke– Smoking is the patient’s responsibility– Nurses who are obviously not models of healthy

behavior are not credible as messengers to adopt them


17

Findings

• Many participants were ambivalent about the measures relating to falls and restraints– Who is responsible when patients fall, the nurse

or the patient?– Don’t some patients need to be restrained?– Do nurses sometimes restrain patients as

“retribution”?– Won’t some patients fall if they aren’t

restrained?


18

Findings

• Measures of nurse-patient staffing ratios, nursing mix and turnover rates, confused many

• Nursing mix: What’s good? Measure doesn’t “point to” the answer

• Nurse-patient ratio: Of great interest, but the measure was just not clear enough to consumers

• “Rate at which nursing staff voluntarily end their employment” – does this mean the turnover rate? the number of vacancies?


19

Findings

• More broadly, these patients are fully aware of the centrality of nurses to their hospital experience

• They realize that in many cases, nurses share responsibility with others, including physicians and patients

• More than anything, people want the nurse to “be there” and to “be competent”

• We can’t assume consumers know the “evidence base” underlying the measures, for example that restraints actually increase rather than reduce falls


20

Implications

• Does this lack of expertise mean we shouldn’t bother asking the public their views? NO!

• The public has much to contribute to assessing quality measures

• Yet at the time of the study, NQF endorsed measures even if there was no evidence of the public’s response to them


21

Implications

• Getting public input before endorsement has these benefits: – Professional judgments about the absolute and

relative importance of individual measures will be either reinforced or seriously questioned

– Issues of attribution of responsibility will be surfaced

– Problems in measure specification that lead to potential confusion can be resolved


22

Implications

• We can get guidance about– When we need to explain the evidence

base to “legitimate” a measure – What kind of “contextual” information

will be needed if measures are to be publicly reported

– We can learn what else the public cares about with respect to a particular issue


23

The Sequel

• NQF always takes measures through a periodic review process; this took place in 2008-09 with the NQF 15

• We were able to present our findings in detail to NQF staff, to the Hospital Quality Alliance, as well as to many nursing and health services researcher

• As a result of this, and the review process, the smoking measures are gone, the measures the public liked most remain, and other measures are being tweaked


24

The Sequel

• NQF has said they are moving toward changing their requirements for measures

• It remains to be seen if this will happen and what it will look like

• The good news: people involved with the review welcomed getting the kind of information we provided, and were surprised at how much it revealed

• The remaining question: no one questions spending money on statistical field testing of measures; will they find the resources to do careful consumer testing as well?

INVOLVING THE PUBLIC IN ASSESSING HEALTH QUALITY MEASURES Shoshanna Sofaer, Presenter Kirsten Firminger, Research Associate.

Documents

public slide

public reports

health care quality

healthcare quality

validity of measures

gauge public response

sequel slide

research associate slide