“Doctors are just the same as lawyers; the only difference is that lawyers merely rob you, whereas doctors rob you and kill you too.” – Anton Chekhov
Quality in health care has become a large driving force for managers, clinicians, and payers. Last week, we explored a pay-for-performance program from Taiwan that demonstrated some of the ill effects of cherry-picking patients when reporting on process and outcome measures.
Researchers in Finland studied the Patient Insurance Centre, which receives almost all injury claims in that country. Annually, nearly 8,000 claims are handled by this administrative consortium (and only 20 annually by the courts) in contrast to the United States’ system. Another important distinction is that the Finnish Patient Insurance Centre can award compensation to an injured patient through a no-fault mechanism. Awards are granted based on one of the following criteria: “treatment injury, infection injury, equipment-related injury, accidental injury, injury from damage to health care facilities, injury due to delivery of pharmaceuticals, and unreasonable injury.”
The researchers investigated over 16,000 first-time hip and 17,000 first-time knee replacements performed in one of 35 public hospitals from 1998 through 2003. Injury claims were compared to already existing and utilized quality indicators: readmission within 14 days after the initial procedure, infection of the artificial joint within 1 year, and surgical revision within 5 years.
Approximately 2.5 percent of each surgical procedure resulted in a claim; 49 percent of hip and 42 percent of knee operations with a claim received compensation.
Analysis of the data detected trends between claims and quality outcomes. For knee replacements claim filings were statistically associated with revisions, infections, readmissions in both unadjusted (Pearson correlation index 0.30, 0.23, 0.19, respectively) and risk-adjusted analyses (Pearson correlation index 0.30, 0.21, 0.20, respectively).
Similarly, for hip replacements, significant associations were evident between claim filings and revisions (Pearson correlation index 0.21) and infections (Pearson correlation index 0.27) in both risk-adjusted and unadjusted analyses.
These significant associations persisted for compensated claims as well as filed claims for hip replacements. However, for knee replacements, claim compensation only remained significantly associated with revisions and infections. Hospital readmissions lacked a significant correlation with claims compensation for knee surgery.
This study merely sought to determine whether injury claims correlated with definable measures of quality care, specifically revision, infection, and readmission following hip and knee replacements.
While statistically significant correlations were uncovered for claims made (and some for claims compensated), these were most robustly true for revisions and infections. Readmissions were slightly less often associated with claims.
An important take-away from this study is that injury claims could be used as a proxy for health care quality in a no-fault system such as Finland. However, in a more litigious society where patient injury is viewed with stigma (as in the United States) caution must precede any introduction of such as scheme into the current quality measures used for hospitals and physicians.
Physicians that tend to see repeated malpractice claims might be delivering lower quality care and by adjusting compensation based on such metrics, payers may be able to prompt improved quality of care. But, the fact that nearly every US physician will be sued in the course of his or her clinical lifetime indicates that medical malpractice claims in the United States might underestimate the quality of care actually delivered.
In American society, a medical malpractice quality metric might simply be a proxy for patient satisfaction not quality of care. Nevertheless, health care payers should study the feasibility of this and all potential process and outcome measures for physician buy-in.