Checking off the box or actually making a difference? How do you know if the language support provided in your HC organization is effective and sufficient? Here is a primer on clinical quality improvement for language access program managers.
Healthcare as a whole is just beginning to assess patient outcomes, which is the true test of effectiveness. Did blood pressure go down? Is the patient managing the colostomy at home successfully? Is the child’s asthma controlled? Patient satisfaction is a very poor proxy for good clinical outcomes, partly because patients do not yet hold HC providers accountable for clinical outcomes, and partly because we never get 100% participation on patient satisfaction surveys. We get very little feedback from language need patients.
Assuming that improved clinical outcomes are desired, how do we drive improvement of clinical outcomes through language support, and how do we get objective proof of outcomes to guide our language support program? We want to keep doing what we are doing well, we want to correct what is ineffective, and we want to start doing what we have not yet worked on, but which needs doing.
Clinical quality improvement focuses on three types of patient groups or processes:
High volume refers to what happens to a lot of patients. If you can fix a high-volume process even a little bit, it will make a big overall impact. For example: How difficult is it for patients to navigate phone menus to reach their care staff? How difficult is it for patients to request medications refills? How many patient intake documents are incomplete because patients cannot read them well enough to fill them out?
Problem-prone refers to specific processes that often go wrong. For example, referred-in patients have to wait until the specialty clinic receives the medical record before they get an appointment. Many patients have trouble arranging to have their medical records sent, and some patients will never be able to present records from earlier years. Care is delayed and, in some cases, patients just give up. Another problem-prone group is patients who need prompt care for serious chronic diseases that flare up: Are they able to get advice from their own care team right away, rather than being shunted to inappropriate ED care by reception staff?
High-risk refers to specific patients for which flawed process can be dangerous or deadly. Patients with symptoms of stroke, for example, must be triaged and receive intervention immediately. There are many conditions for which both the patients and staff must be alert and responsive to signs and symptoms.
Clinical care improvement concepts are also valid for our objective study of how well or poorly our language need patients are faring in our organization. We need to ask how effective our language support program is in driving better clinical care outcomes in high volume situations, problem-prone processes, and high-risk situations. Interpretation and translation are critical to improving care, in conjunction with other quality improvement efforts.
How many language need patients no-show to their appointments? How many referred-in patients ever actually get appointed and turn up for care? How many referred-out patients ever get seen by the outside provider? How many patients seek care in the Emergency Room without contact with their clinic first? How many inpatient days do patients experience with no interpreted encounters? Are children appearing for their vaccinations and well child checks? Are there significant differences between different language communities within these various parameters?
How many of the language need patients who are admitted to the hospital have an Advance Directive on file? Care becomes immensely complicated when an inpatient is no longer able to speak for himself and care decisions are disputed among family members and the care team. An audit can be run on all admitted language need patients to see which ones have an advance directive on file.
How many language need patients show up for procedures without prepped properly at home, necessitating cancelation of the procedure? Do surgical patients fail to arrive fasting? Do colonoscopy patients fail to observe the diet and laxative prep? Do interventional radiology patients fail to modify their anti-coagulation meds? Do dental procedure patients fail to take their prophylactic antibiotics?
Another problem-prone situation involves diabetes nutrition counseling that is not adapted to the patient’s preferred style of cooking and eating. This involves not only language support but cultural adaptation to the patients’ needs. Large resources of staff time, patient time, and insurance money are spent on nutritional counseling. But is blood sugar management improving? Is it improving more in certain language groups than in others? That is the only way to tell whether patients are complying with nutritional guidelines. Consequences are huge. Anecdotal evidence from patients, interpreters, and providers indicates that nutritional counseling is largely ineffective for immigrant patients unless it is culturally adapted.
The C-section rate for Somali women delivering at the hospital may be way higher than expected considering that most Somali women have had multiple births and are low risk. Most Somali women strongly prefer midwife care, but are often assigned physicians for delivery. Somali women strongly prefer vaginal birth over C-section. Due to displeasure with being assigned physician care, with its attendant higher risk of C-section delivery, many Somali women unwittingly set themselves up for the result they are trying to avoid. They often delay in presenting to L and D for a day or more after their water breaks. They then fit the protocol for C-section to prevent stress to the baby. The babies are already stressed because mom waited so long, and they have a higher rate of admission to the neonatal intensive care unit due to meconium aspiration and pneumonia.
Mental illness of certain kinds is high risk. Many language need patients will not accept referral for mental health care because of fear that the community will punish them for any indication of being mentally ill. Some HC organizations have addressed this fear by embedding mental health providers in the medical clinics. No one waiting in the family practice waiting room for the patient knows that the patient has had a visit with a mental health provider.
The HC organization may have trouble finding interpreters to take assignments in high-risk patient situations, such as inpatient mental health care, TB and HIV care, transgender care, and care for congenital deformity in pregnancy . These situations require interpreters who are willing to show up and competent to perform ethically. Having poor language support for these encounters can have steep consequences for the patient and family. The language access program manager should know if the high risk assignments in her HC organization are being filled or not, and whether there are complaints about the interpreter performance from providers or patients.
Data collection and analysis
Data collection and analysis on many of these parameters can be carried out fairly easily. All of the quality reviews suggested above can be carried out without undertaking formal human subject review. Hospitals have processes already established to allow them to run audits on many clinical indicators. Collaborate with the clinical units and IT to add language designators to established reports. Volunteer with various clinical improvement teams to look at specific language communities within their area of focus. The teams will appreciate identifying specific factors that improve care outcomes for language need patients, because their overall statistical average of outcomes will improve.
Patient safety data is a source of valuable information about clinical outcomes that should be mined for process improvement ideas. The patient safety database is populated by staff and providers whenever a patient is put at risk or actually suffers harm. Few if any hospitals run this list of harmed patients against the language designator variable, to identify patients with language need who come to harm. For example, falls that happen in the hospital may be higher for language need patients than for English-speaking patients. Are more non-English-speaking patients discharged after surgery being re-admitted with infections? This might point to un–interpreted discharge teaching or lack of translated discharge instructions that resulted in poor self-care once home.