Home Health and the Puzzle of Hospital Readmissions



Notes from our session at the 2014 New England Home Care Conference


Two of us (Roland Stark and Laurie Courtney) were pleased to present a breakout session, ‘Home Health Services and the Puzzle of Hospital Readmissions,’ at the 2014 New England Home Care Conference. Our presentation was the result of ReInforced Care’s analysis of data from four years’ time and over 120,000 patients. The audience went away informed of some counter-intuitive results, which we call the “Home Health Conundrum,” and engaged in some spirited conversation which will help drive future research. It is our hope that the presentation left attendees thinking about Home Health’s successes; how to measure and promote those successes; whether hospital Readmission Rate (RAR) is an appropriate quality measure of Home Health agencies; and how to carry out their own research on these topics, adding to evidence-based practice in the field.

Whether we agree with this or not, in the current fiscal-centric healthcare environment, RAR is an almost universally used quality measure. There is an ongoing focus on identifying causes of readmissions and in promoting programs that help to decrease those considered avoidable. In our collaborative work with hospitals, ReInforced Care gathers and analyzes a tremendous amount of data. One of the many variables consistently examined is an indicator of Home Health Services (HHS), which includes care from a visiting nurse, physical therapist, occupational therapist, or speech therapist. Every hospital with whom we work has asked us to look into differences between patients who go home with HHS and those who do not.

Here’s the counter-intuitive result. At first pass, we see that patients with HHS readmit to the hospital at a higher rate than those without. Now, no one would suggest that to decrease their RAR hospitals should begin to discharge patients without these services. Individuals who require HHS tend to have complex needs; we are therefore not surprised that they have a higher RAR. But we also expect that if we controlled for these needs, we would see a benefit accrue to the HHS group.

We used various statistical methods of control, such as logistic regression and propensity score matching. For every hospital, the result was a narrowing of the RAR difference between the two groups, but never a reversal, i.e., no sign of a benefit from HHS.

This piqued our curiosity. It didn’t make sense to us. We spoke to others about it – to the hospitals we work with, other colleagues, and friends. We began searching the literature to see what others were saying about this.

And we found ways to dig deeper: we

  • substituted a self-reported HHS measure for a hospital-reported discharge disposition;
  • looked for a difference in the 2- and 6-month RAR;
  • checked for shorter hospital stays on readmission;
  • checked for different results among different HHS agencies;
  • tested for statistical interactions;
  • tried a specialized way to treat missing data – multiple imputation

We even investigated the extremely technical question of whether the accepted statistical method of logistic regression could realistically be expected to show the kind of reversal we sought.

Still, we had no evidence that the HHS group met with better results.

Does this mean that HHS does not reduce readmissions? And if not, is that a negative thing? Patients are discharged from the hospital earlier and sicker than ever before. HHS is the safety net for these patients. It allows patients to return home, rather than stay in the more costly acute care setting. In doing this, HHS is fulfilling its purpose. A return to the hospital may be seen as a positive outcome – a direct result of trained professionals in the home being able to teach patients to recognize and act on red flags.

Is the ‘discharged home without HHS’ the correct comparison group? HHS pts may be fundamentally different from non-HHS patients and therefore a better comparison group might include those discharged to a skilled-nursing or long-term-care facility. Patients on ventilators, with wound VAC treatment or Pleur-evac® chest tubes, with multiple comorbidities, the frail and homebound, etc. usually go home with HHS or to another facility. They are not generally discharged home without any support services, and therefore it may not be appropriate to measure their outcomes against those of other patients who go home.

Is RAR an appropriate measure of the quality and success of Home Health agencies? RAR is a measure that has been somewhat thrust on Home Health. It is interesting that the benefits of Home Care in perinatal and maternal-child populations are well documented by groups such as the Nurse-Family Partnership and Home Visiting Evidence of Effectiveness (HomVEE), and that the documented benefits include measures of health, injury, childhood development, school readiness, and socio-economic status, but do not include hospital RAR. Home Health agencies may need to shift the focus to measures that better showcase their strengths and successes. Mortality, chronic disease or disability management, physical or mental functioning, and quality of life have been suggested as quality measures better suited to HHS.

Related studies published in the literature include multiple care-transition and post-discharge interventions, and the results are extremely mixed. We hope that our findings add to the conversation, help HHS providers to more accurately and affirmatively proclaim their worth, and spark an interest in others to conduct their own research in this area and to share their results.

Back to Blog & Resources