1. Will the patient satisfaction system meet the criteria for accreditation or qualification?

Arbor's Patient Expectation Project has been used by other PEP clients to meet JCAHO and AAAHC requirements. Arbor also has developed, tested and implemented the HEDIS survey instrument to measure satisfaction of health plan enrollees in order to comply with NCQA criteria. Likewise, we are happy to add questions to be asked of all patients/users/enrollees which will assist in applying for Baldrige Awards.

Back to Top

2. Data must be comparable within our network of services. Will we be able to compare the satisfaction levels of hospital patients with the satisfaction level of physician office patients and home care patients?

The survey instrument is designed for use at three levels of the organization. Therefore, input will be obtained from the people who will use the survey results at each level. This will insure that statistical measures of patient satisfaction and verbatim patient comments are focused and actionable. Survey questions will fall into one of the following categories:

• System Core Questions: These questions will cover issues of concern to all patients/users/enrollees, no matter which of the system's services they may have used. Issues include such things as perceived staff competency and compassion, adequacy of information/education received, confidentiality and trust. A series of demographic questions will also be asked in order to permit cross-referencing. A composite rating of network-wide satisfaction will also be developed and tracked.

The results of the core questions will allow executives working at the system level to measure and track overall performance across all provider sites on a uniform set of issues known to matter most to patients and their families and to track the overall performance of the system over time. Comparative information from other systems will also be provided.

• Service-Specific Questions: The survey instrument will contain "modules" of questions covering issues of concern to specific groups of patients. For example, the emergency module focuses on the various elements of waiting time and keeping family members informed. Modules have already been developed and can readily be applied for doctors' offices, home care patients and many other services. A service-specific module will be asked of all patients receiving that service at any of the system provider sites. The results of these questions will allow system managers to compare service-specific patient satisfaction across all provider sites using a consistent measurement tool which is sensitive to the concerns unique to different patient groups. Service-specific results may also be useful in marketing the ability of the system's providers to deliver a consistent level of patient satisfaction wherever care is offered.

• Provider-Specific Questions: In order to make the most of the opportunity for significant patient feedback, the third level of survey design will permit service managers at individual sites to ask questions of unique concern to them. For example, if the manager of one of the physician offices is considering remodeling the waiting room we can insert questions on patient preference. Once the issue has been addressed, these questions can be deleted from the survey and others substituted as necessary.

Back to Top

3. Will the data be comparable with other organizations so that we can benchmark and set goals for improvement?

Arbor currently provides comparative in three ways: internally across the hospital's services, statewide, and regionally. The hospital is able to select members of its comparative group and from among other Arbor clients. It is possible to compare satisfaction ratings at the service-specific level with other hospitals and other integrated delivery systems. Arbor intends to provide national data from H-CAHPS as soon as data is available.

Back to Top

4. How and by whom will assistance, such as the statistical significance of changes or interpretation and recommendations, be provided and are these services included in the pricing?

Hospital staff will not be expected to perform tasks such as mailings, reminders, transcribing patient comments etc. Instead, a regular reporting system will be established with the data processing department to electronically provide Arbor with information about patients on a regular basis. Arbor randomly selects patients from these files for interview. No other tasks or expenses are required of hospital personnel to maintain the project.

Arbor will make an on-site Power Point presentation of each report and provide extensive discussion and interpretation. Also, examples of what other hospitals have done to address similar issues will be cited. Contact people and telephone numbers of other Patient Expectation Project clients are readily available for inter-hospital networking. In addition, Arbor hosts a Patient Expectation Project Client Seminar each fall at which many useful presentations and discussions take place between new and experienced PEP users.

Back to Top

5. Will the data be statistically valid and reliable in order to have credibility with employers and physicians?

We feel most confident using telephone interviews conducted by our trained staff to avoid unintentional bias and to maximize accuracy. Because our experience shows that 88% - 92% of the patients we contact will complete the interview, we can virtually eliminate the danger of sampling bias that comes with a written questionnaire where response rates are typically 15% - 25%.

The quality of Arbor's telephone interviews allows us to make extensive use of open-ended questions throughout the survey. We find that this type of qualitative question is most productive as a follow-up to a quantitative question. Open-ended responses are recorded and reported verbatim in order to give the "feeling" behind the numbers.

Strict adherence to random sampling methodology, the distribution of interviews over several-month intervals, and rigorous statistical analysis assure the credibility of the PEP survey results.

Back to Top

6. Does the analysis provide predictive indicators to help us focus our efforts on where they can have the greatest impact?

A negative response to one or more of a panel of four key indicators will result in a patient being classified as dissatisfied with care overall. The hospital-wide percentage of satisfied/dissatisfied patients and the percentage in each service will be calculated for each report. Internal baseline levels of satisfaction will be established for each service to be studied. These baselines will be used to trend internal, service-specific changes in patient satisfaction over time.

The survey instrument will cover all aspects of patients' experiences. Multiple regression analysis of the responses of the "dissatisfied" patients will enable Arbor to pick out the key elements which most directly impact overall satisfaction/ dissatisfaction. These are the "hot buttons" which, when they go well in a patient's experience, can overcome lesser frustrations or disappointments. These are also the things that can wipe out all the positive things that may have happened if they go wrong. Managers can reinforce the things staff is doing which build satisfaction and loyalty. The "dissatisfiers" become the focus for quality improvement teams to work on. Subsequent survey iterations will measure the success of the efforts.

Back to Top

7. There is a recognized balance to be achieved between national comparability and customization. How does your company achieve this balance? To what extent are the questionnaires customized and how is this done?

Each project is different. Each is tailored to the needs of the client and changes as those needs change. Yet, survey instruments have a lot in common across clients because their patients do. In this way, we can ensure extensive comparability of results.

At the beginning of the project, considerable effort will be made to "customize" the survey instrument to the information needs of executives and the managers of each service to be studied. Nevertheless, we strongly encourage the continual revision of the survey instrument. Additional questions can be inserted on a service-specific or organization-wide basis at any time. In addition, we will review the survey once a year to suggest improvements to question formats, new issues to look into and any unproductive questions that can be dropped. There is never any charge for revising the survey.

Through focus group discussions, patients have been involved in the design of the survey instrument for all of the service-specific modules and "bookend" questions. Arbor also offers the option of conducting additional patient focus groups as an added source of the input into the survey instrument.

Back to Top

8. Generally speaking, how would you recommend we proceed with this project? Who do you suggest we survey and how often? Are there important issues/questions we should raise?

The overall purpose of the Patient Expectation Project is to reinforce the things the hospital is doing that create satisfaction and build loyalty and to identify and change the things patients don't like.

Most would agree that people and institutions don't change behaviors on a monthly or even quarterly basis. Formal reports compiled too frequently will not give enough time for statistically significant changes to take place. Since real change rarely can be measured after every month or quarter, overly frequent reporting is not cost-effective.

We generally recommend that interviews be spread over a 3-month period to avoid unusual or seasonal factors. For most services a frequency of two formal reports per year is recommended. By distributing the calls evenly over several months, we can avoid the impact of specific short-lived events such as a pediatric flu epidemic in January or a rash of sports injuries in July. By reporting two times per year rather than four, we can reduce the cost of the PEP without sacrificing accuracy, sensitivity or utility.

While the formal, detailed statistical report along with the comparative information is usually prepared and presented twice each year, all patient comments to open-ended questions are available to the hospital through weekly, bi-weekly, or monthly updates to ArborOnLine. In fact, Arbor keeps patients in touch with providers in three ways.

1. Daily notifications of any survey respondents requesting a call to discuss a concern.
2. Weekly, bi-weekly, or monthly updates to ArborOnLine providing numerical reports and verbatim patient comments.
3. Semi-annually formal, detailed written reports and presentations will be delivered on-site. While it is certainly possible to separately survey each category of service at each provider site, that would be neither cost-effective nor worthwhile. Arbor's experience in this field is that meeting or exceeding patient expectations is more difficult for some services than others. On the inpatient side for example, the parents of pediatric patients are often most anxious, want to be involved in their child's care and are generally more difficult to support. Similarly on the outpatient side, emergency patients are difficult to please because they are intolerant of waiting times and often highly stressed. For these types of patients, we recommend surveying at the service-specific level in order to keep patients in close touch with providers.

Other services which typically are low-stress, such as routine outpatient radiology or laboratory testing, can be surveyed as a single category. There is no need to expend the resources required to survey each modality separately unless special circumstances so indicate. Of course, special initiatives such as the offering of a new service or the marketing of an existing one may dictate service-specific surveying for other purposes.

The second design consideration is how often to repeat the survey within a year. Services known to be difficult or to have volatile patient satisfaction or ones where a "before and after" measurement is needed should be surveyed at least two times per year. In contrast, other services with low volume or a stable patient group such as home care, skilled nursing or physician office practices need only be surveyed once per year. For these services, one survey per year will suffice for internal management purposes and for accreditation requirements.

Back to Top

9. Will patients be surveyed in a timely manner after the use of a service?

The PEP survey instrument is a detailed chronological review of the many elements of a patient's experience. Therefore, it is necessary to contact patients soon after going home when recall is at its highest. The preferred method of conducting the telephone interview is for the data processing department to provide Arbor with a computer file after the close of each month containing information on patients who used the services included in the survey during that month. The hospital is encouraged to add nursing unit, DRG, ICD-9-CM codes, physician identifiers, or any other information to the record which can later be used for cross-references. Four attempts - two weekday and two evening/weekend - will be made to complete the random sample of interviews.

Under Arbor "QuickTurn Service" the hospital may provide us with data as frequently as they are able. Arbor will update data and patient comments to ArborOnLine on that same schedule. For example, if we receive patient data on Wednesdays, then by the following Wednesday patient comments and survey counts for that week will be available.

Back to Top

10. How are patient complaints and issues needing immediate follow-up handled?

Each telephone interview begins by assuring respondents that their answers are anonymous. Each interview will conclude by asking the respondent if they have any unresolved issues that they wish to have someone from the hospital call them about. If so, they will then be explicitly asked for permission to give their name and telephone number to the hospital for follow-up. They will also be asked to briefly describe the nature of the issue (donations, billing, personnel, lost and found, etc.) so the appropriate person can prepare for the call. The names, numbers and nature of the issues will be faxed or e-mailed to a pre-designated person at the hospital that same day for daytime interviews, or the next business day for evening or weekend interviews. Modifications to this methodology can be made to facilitate its use for risk management, if desired.

Back to Top