Relating focus to quality and cost in a healthcare setting

Abstract

Focus in healthcare has been heralded as the next frontier in improving its efficiency and efficacy (Herzlinger 2004). Focus takes several different forms, ranging from standalone specialty centers to a hospital that places a strategic emphasis on a clinical area. We adopt this latter perspective and define focus as a disproportionate emphasis on a particular clinical area in a hospital. We use secondary data from hospitals providing cardiology care in New York State to examine the relationship between focus and performance. We develop two measures of focus. Proportional focus is defined to be the proportion of cases treated in a particular clinical specialty. Expertise focus is defined to be specific evidence that a hospital has taken action to build expertise in treating diseases in that specialty. We operationalize hospital performance along cost and quality dimensions, and we use hierarchical regression to examine the impact of focus on performance. Our results indicate that proportional focus, but not expertise focus, is associated with better cost performance. Quality performance, on the other hand, was associated only with the interaction between proportional focus and expertise focus, which means that only hospitals exhibiting higher levels of both proportional and expertise focus achieve better quality performance. These findings support the notion that not only is focus important in healthcare, but also that researchers and practitioners need to recognize that relationships are contingent on the performance and focus measures used and thus, findings may not be generalizable from one metric to another.

Appendix. Example Calculation of Percentage Difference from Expected Mortality Rate

For one hospital, for APR-DRG = 190 (acute myocardial infarction) the expected mortality rate is computed in the following manner:

Risk Subclass	Cases	Deaths	Subclass State Average Mortality Rate	Weight	Subclass Weighted Average Mortality Rate
1	21	0	0.0034	21/137 = 0.153	0.153*.0034 = 0.0005
2	43	1	0.0184	0.314	0.0058
3	40	4	0.1120	0.292	0.0327
4	33	13	0.4172	0.241	0.1005
Totals	137	18			0.1395

Individual weights are calculated by dividing the number of cases in a subclass by the total number of cases, so for the risk subclass 1, the weight would be calculated as:

$$ {\text{Subclass}}\,1\,{\text{weight}} = 21/137 = 0.153 $$

The weighted average mortality rate for each subclass is calculated by multiplying the state average by the weight, so for subclass 1, the weighted average would be calculated as:

$$ {\text{Subclass}}\,1\,{\text{weighted}}\,{\text{average}}\,{\text{mortality}}\,{\text{rate}} = 0.153*0.0034 = 0.0005 $$

The expected mortality rate is then calculated by summing the individual subclass weighted average values. In this example, the expected mortality rate:

$$ {\text{Expected}}\,{\text{mortality}}\,{\text{rate}} = 0.0005 + 0.0058 + 0.0327 + 0.1005 = 0.1395 $$

The observed mortality rate is simply the total number of deaths divided by the total number of cases:

$$ {\text{Observed}}\,{\text{mortality}}\,{\text{rate}} = 18/137 = 0.1314 $$

The mortality rate performance variable (percent difference from expected mortality rate) is then calculated as follows:

$$ {\text{Percent}}\,{\text{difference}}\,{\text{from}}\,{\text{expected}}\,{\text{mortality}}\,{\text{rate}} = \left( {{\text{Observed}}\,{\text{mortality}}\,{\text{rate}} - {\text{Expected}}\,{\text{mortality}}\,{\text{rate}}} \right)/{\text{Expected}}\,{\text{mortality}}\,{\text{rate}} $$

In this example, the percent difference from expected mortality rate would be calculated as:

$$ {\text{Percent}}\,{\text{difference}}\,{\text{from}}\,{\text{expected}}\,{\text{mortality}}\,{\text{rate}}\,\left( {\prime \prime {\text{risk}}\,{\text{adjusted}}\,{\text{difference}}\,{\text{in}}\,{\text{mortality}}\prime \prime } \right) = \left( {0.1314 - 0.1395} \right)/0.1395 = - 0.0581*100\% = - 5.81\% $$

This hospital’s actual observed mortality rate for this APR-DRG is therefore 5.81% lower than the mortality rate expected given its mix of cases. Negative values indicate an observed mortality rate that is lower than expected; consequently lower values indicate better performance. The overall expected and observed mortality rates for a hospital are calculated by aggregating expected and observed mortality rates, respectively, for all APR-DRGs, and the overall percentage difference from expected mortality rate is calculated by using the overall observed mortality rate and overall expected mortality rate in the formula above. Similarly, we use the percent difference from expected costs as a measure of the cost performance for the hospital. We compute the risk adjusted difference in cost measure as $\left( {{\text{Observed}}\,{\text{costs}} - {\text{Expected}}\,{\text{costs}}} \right)/\left( {{\text{Expected}}\,{\text{costs}}} \right)$.