Scott MacStravic, PhD
With most hospitals and other HCOs barely able to keep enough staff to meet quality standards, and often having to rely on involuntary overtime even to do that, the challenge of attracting, motivating, and retaining employees has never been greater. The concept of “talent management,” i.e. identifying, attracting, managing, motivating and retaining truly talented individuals, in contrast to merely adequate staff, may seem far beyond what job market and revenue realities permit.
But HCOs that can succeed in boosting the talent levels, as well as sheer numbers of staff they have compared to what they need, will enjoy significant advantages in their market. First, their more talented staff will, by definition, deliver a higher level of quality, customer service, word-of-mouth “advertising” impact, efficiency, etc. and thereby improve the HCO’s performance. As pay-for-performance (P4P) systems grow, and with them the proportion of the HCO’s revenue that depends on better performance, the direct revenue impact of having higher talent levels will increase, as well as the indirect.
HCOs have the same choices in evaluating and compensating talent as they do with respect to evaluating the quality of care they deliver – based on any mix of “structure, process, outcome and value” metrics. [A. Donabedian The Definition of Quality and Approaches to Its Assessment Ann Arbor, MI, Health Administration Press 1980] While HCOs have long relied mainly on structure and process measures in their own management, and marketing for that matter, it makes sense in today’s market, to move more to outcome and value.
Of course, the extent to which these measures will work better depends on the extent to which the HCO’s revenue and operating margin can be linked to them, in contrast to structure and process measures. If P4P systems lean more toward structure measures, such as whether the HCO has CPOE and EMR systems in place, board-certified physicians and enough RNs, for example, then these measures become more important. If such systems lean toward measures of adherence to EBM standards, HEDIS gauges of doing the right thing, etc., then process measures become more important.
Increasingly, however, the measures that are being collected for P4P systems, and reported in “scorecards” made public for guiding payers’ and consumers’ selection of providers, are leaning toward outcomes, such as mortality and complication rates and patient satisfaction levels. And there is definitely a move toward value measures, as well, such as cost savings achieved through case, care, and disease management.
When HCOs are paid on the basis of explicit measures of structure, process, outcomes, or value, it makes sense for them to measure their talent in precisely the same terms as it is being paid. And once measured, it makes sense for them to pay their talent based on the same measures. This can enable achieving the universally touted alignment of incentives between the organization and its workforce.
But it will require throwing out one of the more popular devices currently used in evaluating and paying talent – the “normal distribution”.
In virtually all of the HCOs where I was employed during my thirty-year career in healthcare, the “bell curve” was used in judging and compensating employees. The organizations’ overall distribution of performance ratings, as well as compensation levels, was expected, and therefore forced to fit a distribution where “excellent” was reserved for a small minority of staff, usually around five percent or less, with “good” permitted for perhaps another ten or fifteen percent, “average” expected for fifty percent or more, “fair” for ten or fifteen, and “poor” for five.
This distribution was used to examine the ratings actually rendered by managers, and any who were thought to be “over-rating” their employees were advised to revise their evaluations. Raises were limited to cost-of-living or perhaps one percent above that for all but the top ten or twenty percent of performers. And thanks to the symmetry of the normal distribution, raises for those forced into the fair and poor categories were less than cost-of-living levels to make sure total compensation was “budget neutral”.
I recall the problems I had when I was corporate VP of Strategy and Marketing, and rated all four of my key professionals and managers in the good to excellent range, with bonuses for all four. I argued that all were truly significantly better than average, and that since our division had produced a significant revenue and margin addition for the system, all should be eligible for such a bonus. While I won the battle that year, I lost the war, and all four were removed from the bonus system the next year.
The only people on my staff who were exempt from this normal distribution tyranny were the physician relations representatives. They were compensated through bonuses directly tied to the number of additional admissions, and amount of revenue and margin delivered thereby, from physicians they were visiting and “developing” into better customers. And the formula for calculating their annual bonus ensured that their total compensation was dramatically affected by this bonus, compared to their salary. This ensured that all who were good to excellent performers got good to excellent compensation for it. It also ensured that they would remain among the organization’s talent.
While there is some justification for enforcing a normal distribution on managers’ ratings of their staff, it makes no sense once a value performance measurement and management system is in place. As soon as the value delivered by employees can be measured, this value should determine how much they are paid, not an artificial statistical distribution. And the conscious, public aim of the HCO should be to produce a staff with as high a proportion of “talent” rather than “placeholders” as possible, which will only happen if all talent is evaluated and compensated accordingly.
Both the level of compensation and the proportion of total compensation that is based on performance have been identified as major factors in attracting, motivating, and retaining talent. [[W. Lynch, et al. “Brief 1: Human Capital Motivation and Productivity” Health as Human Capital Foundation May 2007 (www.hhcfoundation.org/hhcf/pdf/Brief1.pdf)] The more effectively these are used, the more the HCO’s workforce should become “biased” in the direction of high performers, ideally reaching the Lake Wobegon ideal of everyone being above average.
One classic example of this effect occurred in the first year that one firm switched from a normally distributed hourly wage compensation system to one based on objectively measured productivity. The major direct result was a 44% improvement in productivity at the cost of only a 10% increase in compensation. But perhaps an even better result was the reduction in turnover by 21% among those rated as high-performers, while turnover among those rated average or low performers increased by over 10%. [E. Lazar “Performance Pay and Productivity” American Economic Review 190:5 Dec 2000 1346-1361]
It is a simple mathematical task to calculate how the mix of “talent” in the overall workforce will change away from the normal under such circumstances. As turnover decreases among high-performers and increases among average and low performers, the average performance levels will improve significantly. As the organization improves over time in identifying potential talent in its new hires, by learning what distinguishes high performers and can be used in applicant screening, this move toward an “abnormal” distribution in terms of talent levels will increase.
It should be the challenge of C-level executives, certainly the CEO and COOs, to deliberately seek and strive for such an abnormal distribution. One of the best ways to do so will be to ensure that the organization’s P4P system permits it, since otherwise, forcing a normal distribution in payment will end up forcing a normal distribution in performance. And a normal distribution in performance will automatically force a normal level of payment to the HCO, which may not be enough to enable it to succeed, or even survive in future.
