Effect of pay for performance on the management and outcomes of hypertension in the United Kingdom: interrupted time series study
BMJ 2011; 342 doi: https://doi.org/10.1136/bmj.d108 (Published 25 January 2011) Cite this as: BMJ 2011;342:d108All rapid responses
Rapid responses are electronic comments to the editor. They enable our users to debate issues raised in articles published on bmj.com. A rapid response is first posted online. If you need the URL (web address) of an individual response, simply click on the response headline and copy the URL from the browser window. A proportion of responses will, after editing, be published online and in the print journal as letters, which are indexed in PubMed. Rapid responses are not indexed in PubMed and they are not journal articles. The BMJ reserves the right to remove responses which are being wilfully misrepresented as published articles or when it is brought to our attention that a response spreads misinformation.
From March 2022, the word limit for rapid responses will be 600 words not including references and author details. We will no longer post responses that exceed this limit.
The word limit for letters selected from posted responses remains 300 words.
In 1990 the then new GMS contract came into effect, and one of the
requirements was to record the blood pressure, BMI, smoking status and
alcohol intake of 90% of the practice population between 45 and 74 in the
previous 5 years.
There was no requirement to do anything about it and the he minimal
funding came from the total amount available (The Red Book funding formula
for the whole of general practice was"(Intended average income for a GP +
estimated expenses)x number of GP principals - and this was then split
into fees and allowances to produce maximum work for minimum expense to
the treasury: any 'overperformance' being rewarded by it being clawed back
the following year.)
Once a raised blood pressure had been detected, I suspect that the
majority of practices treated it.
Of course the other effect of Kenneth Clarke's "new contract" was to
force all practices of any size to invest in computer systems - and again,
the 50% funding was removed from the pool previously supporting premises
and 70% of staff budget (the practice staff budget was frozen at the
levels on 31st March 1990)
In these circumstances, is it surprising that there appears to be
very little difference between blood pressure control before and after the
introduction of QOF?
Why chose hypertension? I suspect diabetes would be a better choice: I
spent a lot of the first year getting information out of secondary care -
and yes, diabetic care did improve as a result!
Competing interests: No competing interests
The recent article cited in the British Medical Journal entitled "The
Effect of pay-for-performance on management and outcomes of
hypertension...." Highlights the limits of an incentives based solution to
the healthcare dilemma faced by the entire world. It is not a lack of
incentives or compensation, but a fundamental economic misalignment of
payment.
We at the Western Clinician's Network (a non-profit organization) are
pursuing research to understand the essential improvements in the
compensation relationship that will allow for pure pay-for-performance. It
is only through this type of revolutionary approach that we can unleash
the ingenuity, and resourcefulness that serves free markets all over the
world. The challenge is in establishing a method of estimating the
relative difficulty of aiding patients to a measurably healthier future.
The BMJ article concludes that the reason no significant improvements
in the measures are discovered was because the "study suggests that care
for hypertension in the United Kingdom was already close (or along the
way) to reaching the threshold required to achieve maximum payments set in
the pay for performance policy " and "The setting of the indicator
thresholds for maximum payment close to prevailing practice may have
provided little incentive for further improvement. Thus, pay for
performance may have simply supported existing practice for hypertension."
We have found in our research at the WCN that this is indeed the case, but
perhaps even more enlightening is our finding that those that can and will
easily change their medical course are already on their way. Those that
require greater levels of intervention or assistance will not be addressed
until the economic drivers are properly aligned. We believe this can only
happen through the equivalent of an economic revolution.
The first step required to achieve this revolution is to create a
meaningful picture that both provider and patients can easily grasp and
use as a guide. So called quality and outcome measures are highly complex
pictures of the adequacy of care, and have little direct impact on
provider or patient behaviors. In two pilot studies we have used a simple
predictive analysis, using the Framingham Risk Calculator which is
publicly available, to consolidate abstract measures into a simple
statement of the probability of a cardiovascular event in the coming 10
years. A more sophisticated resource is available through Archimedes
Indigo product and has been shown in their research to significantly
improve patient compliance with treatment interventions. These and other
innovations cannot currently be incorporated in primary care practices
because there is no economic model that allows their purchase and support.
This would not be the case if providers were paid for measureable
improvements in the health of the individual
The next essential component of compensation redesigned to pay for
measureable improvements in health of patients is to estimate the
probability that a person will choose a healthier future.
When the insurance and HMO industries consider this from their perspective
it equates to a loss avoidance plan and so the patients are "risk
stratified" as an estimate of the likely medical costs to come in the
manageable future. This however continues to force the medical profession
into a cost containment mentality that is at ethical and practical odds
with the profession. Aligning compensation with the ethical roots of
medicine however is now possible.
The probability that a person will choose a healthier future is
dependent on a variety of factors. In order to define the probability of a
person choosing a healthier future, we are seeking correlations within the
demographic and economic measures readily available in electronic health
records and billing data. While these factors may suggest a probability of
a person being able to choose and execute the behaviors required to
realize a healthier future they do not expressly say anything about the
individual. In fact the laws of probability very explicitly exclude this
type of conclusion. The estimation does however allow for a more accurate
means of estimating those demographic and economic factors which will lead
to success in a therapeutic relationship.
Let's imagine that we now have a means of accurately predicting who
will be able to not only choose, but also execute a plan for a healthier
future. This then becomes the last essential component required to
redefine compensation for healthcare: payment for measureable improvements
in outcomes. We will then see compensation go up proportionately for those
patients who require greater assistance to realize that healthier future.
This also frees the provider of an encounters based financial model and
encourages the incorporation of a wider array of resources to achieve this
end, many of which are still being invented. The economy now turns on the
same drivers of performance that the clothing industry (or any number of
other vendors in the free market) does, and produces a higher quality
product at decreasing costs. The need for micro regulatory dabbling is
also altered, since it is in the end a payment for improved health with a
known and visible outcome: decreasing risks of disease and the concomitant
costs and pain and suffering.
# # #
Carl Heard, MD, MMM
Carson City, NV 89703
carlheard@carlheard.com
Competing interests: No competing interests
Reading this and other articles concerning the QOF concept and
outcomes, I am not surprised that improvement in quality of care has been
a little disappointing. Quoting McWhinney: "In a complex system, cause and
effect are not usually close to each other in time and space, and since
organic processes are maintained or changed by multiple influences, it is
difficult to predict the consequences of an intervention"[1]. In
hospitals, where events are close together and descontextualised, process
and outcomes may be more easily measured, but in primary care it just
seems not to be the case.
The RCT creates artificiality. GPs work in a complex environment with
no exclusion criteria, no drop outs, no blinding, but rather a face-to-
face reality full of uncertainty.
On the other hand, the current scientific-bureaucratic medicine -
that translates research evidence into 'clinical guidelines' and now links
this to incentives - may have a reductionist effect on GPs' role, despite
a discourse of 'no change' [2]. Experts in economics and sociology have
described the 'crowded out' effect whereby monetary incentive has the
potential to impair self-determination. They also recognise that this is
less likely to occur in more mechanical tasks than in creative ones [3].
If the QOF prompts the 'no change' perception this creates the rhetoric
paradox: where a holistic approach - which needs creativity - can be
mechanised and reduced to agreed targets (as no 'crowded out' effect was
found). How this new profile of GPs' profession can still be included in
McWhinney's statement of what makes GPs different: "Other fields define
themselves in terms of content: diseases, organ systems or technologies.
Clinicians in other fields form relationships with patients, but in
general practice, the relationship is usually prior to content".
80% of what is considered as quality of care in the QOF relates to
clinical and organizational domains leaving around 20% for the patient's
experience domain and depth of quality measurement. Are GPs content
changing from patient- to target-centred medicine? Should GPs be redefined
as focal specialists on agreed targets?
I am a GP from Brazil currently studying medical anthropology at
Durham University in the UK. I would like to understand the new context
that QOF creates in GPs' workplaces in the UK. It seems that the audit
culture and the new managerialism is transforming the profession of
General Practice for the worse; similar trends are now in play in Brazil.
[1] McWhinnney, I. The importance of being different. British Journal
of General Practice, 1996, 46, 433-436.
[2] Checkland, K. Harrison, S. The impact of the quality and outcomes
framework on practice organisation and service delivery: summary of
evidence from two qualitative studies.Quality in Primary Care 2010;18:139-
46
[3] Marshall, M. Harrison, S. It's about more than money: financial
incentives and internal motivation. Qual Saf Care 2005; 14:4-5. doi:
10.1136/qshc.2004.013193
Competing interests: No competing interests
While Serumaga et al(l) claim to demonstrate the failure of pay for
performance in the UK to improve hypertension, their analytic method is
seriously flawed. The authors use a cross-sectional approach for blood
pressure evaluation, establishing each quarter the percent of tested
patient's in that quarter achieving the desired target and then trending
that proportion across time. They fail to use a cohort approach which
tracks a defined group of patients across time. As a result they may
suffer the well known "clinician's illusion"(2), the bias that comes with
prevalence biased sampling - those patients who are the sickest are
overrepresented in the clinical sample. In our case, those patients under
worste control would be more likely to be retested. Even though
patients might be brought under better control from the cohort
perspective, aggressive testing in those who are not under control would
falsely under-represent the success.
In our work at Montefiore, we have developed software, Clinical
Looking Glass(3, 4), that allows practitioners to create cohorts and
evaluate the outcomes by cohort not by cross-sectional view.
I have little doubt that the author's core conclusions are correct. They
describe no process of informational feedback to the managers of patient
cohorts, no active process of remediation for those patients identified as
failing, and it is hard to imagine that nonspecific dollars without
purposeful augmentation of process could accomplish the desired quality
goals. However, as we begin to seriously evaluate health care delivery
goals we should standardize on the cohort not cross-sectional view.
Eran Bellin, M.D.
Professsor of Clinical Epidemiology and Medicine
Albert Einstein College of Medicine
Montefiore Medical Center
Bronx, N.Y.
Reference List
1. Serumaga B, Ross-Degnan D, Avery AJ et al. Effect of pay for
performance on the management and outcomes of hypertension in the United
Kingdom: interrupted time series study. BMJ 2011;342:d108.
2. Cohen P, Cohen J. The clinician's illusion. Arch Gen Psychiatry
1984;41(12):1178-1182.
3. Clinical Looking Glas. 1-28-0011.
http://exploreclg.montefiore.org
4. Bellin E, Fletcher DD, Geberer N, Islam S, Srivastava N.
Democratizing information creation from health care data for quality
improvement, research, and education-the Montefiore Medical Center
Experience. Acad Med 2010;85(8):1362-1368.
Competing interests: No competing interests
Hypertension is higly prevalent. Hypertension is related to some
severe complication. Good quality of hypertension services is needed for
preventing its complications. The high prevalence of hypertension in the
adult population and the frequent problem of inadequately controlled blood
pressure make a strong case for implementing effective strategies for
improving the care of hypertensive patients.High-quality care for
hypertension includes: Awareness of preventive care for high blood
pressure, regular blood-pressure screening,involvement of multiple
clinical specialties,effective communication between health care providers
and their patient, and active self-management by patients.
This study
showed that pay for performance initiative had no effect on processes of
care and outcomes for patients with hypertension. Cooperation between
patients and providers is needed for better blood pressure control.
Competing interests: No competing interests
Blood Pressure is falling over Time
Serumaga et al(1) report that the Quality and Outcomes Framework
(QQF) has had a minimal impact on hypertension care. While we broadly
agree with this conclusion, we are concerned about their analytic
approach, which we believe may have failed to address key issues.
We were surprised that their analyses failed to show significant
declines in adult blood pressure (BP) levels over time (Figure 1 showed a
slight increase). A long-term trend of steadily falling BP levels has been
observed in Health Survey for England data for a sample of the general
population(2), and similarly across Europe(3). We are concerned that this
may reflect the changing characteristics of the different individuals
included in different time points (quarters) of their analysis. By
contrast, in another large UK primary care database (DIN), we reported
falling BP in the same hypertensive patients followed over time between
2001 and 2005(4).
The rising blood pressure levels in their study likely contributed to
the decline in "controlled" blood pressure they also reported (Figure 2).
This again contrasted with our analyses in DIN(4), as well as a published
report from QRESEARCH(5), which both demonstrated strong time trends
between 2001 and 2006 for a higher proportion of measured hypertensive
patients with BP measurements under the QOF target (150/90mmHg). Although
we agree with Serumaga that the quality of care for hypertension was
improving, their Figure 2 appears to contradict this conclusion.
We believe their methodology in assessing the effectiveness of QOF
has two serious flaws. Firstly they collected data in quarterly
increments, whereas the QOF target that GPs will be working towards is a
measurement (preferably below 150/90mmHg) in the 9 months before the first
day of April each year. It is likely that a patient who has a measurement
that meets the target level in any of the three quarters before the end of
each QOF assessment period may have no further measurements made in that
period. Data collection in quarters suggests a failure to appreciate the
way that QOF is implemented in practice and may lead to unintended biases
(e.g. patients over the target are more likely to be re-measured).
Secondly we do not believe their overall interpretation that incremental
changes towards achievement of a target are automatically of equal value.
For example, moving from 60% to 70% of patients making the QOF BP target
is not necessarily equivalent to moving from 70% to 80%. As one approaches
100% the patients failing the target will be more atypical, and will
represent a greater treatment challenge. Thus we believe it is premature
to conclude that the trend in improved achievement of target BP seen post
QOF was unaffected by the implementation of QOF targets, as it may be
that, without QOF, the trend of target achievement for these more
"difficult" patients may have levelled off.
Finally, their claim that there was "no evidence of gaming of the
system to achieve quality targets" was in contrast to recent findings we
reported on the impact of QOF on blood pressure levels(4). We found that
post QOF, GPs were more likely to assign values just below the target
(e.g. 148, 149), and this has exaggerated blood pressure falls post QOF
(the true prevalence of systolic BP >150mmHg in 2004-5 may be around 4%
higher than recorded data suggest). Importantly however, we found no
evidence of adverse effects of this recording bias on clinical management.
1. Serumaga B, Ross-Degnan D, Avery AJ et al. Effect of pay for
performance on the management and outcomes of hypertension in the United
Kingdom: interrupted time series study. BMJ 2011;342:d108.
2. Health Survey for England. Trends in blood pressure levels.
http://www.heartstats.org/datapage.asp?id=1000, Last accessed: 9/2/11.
3. Danaei G, Finucane MM, Lin JK, Singh GM, Paciorek CJ, Cowan, MJ,
Farzadfar F, Stevens GA, Lim SS, Riley LM, Ezzati M. National, regional,
and global trends in systolic blood pressure since 1980: systematic
analysis of health examination surveys and epidemiological studies with
786 country-years and 5*4 million participants. The Lancet 2011.
doi:10.1016/S0140-6736(10)62036-3
4. Carey IM, Nightingale CM, DeWilde S, Harris T, Whincup PH, Cook
DG. Blood pressure recording bias during a period when the Quality and
Outcomes Framework was introduced. J Hum Hypertens 2009; 23(11):764-770
5. Time Series Analysis for selected clinical indicators from the
Quality and Outcomes Framework 2001-2006. http://www.ic.nhs.uk/statistics-
and-data-collections/audits-and-performance/the-quality-and-outcomes-
framework/time-series-analysis-for-2001-2006-for-selected-clinical-
indicators-from-the-quality-and-outcomes-framework Last accessed: 9/2/11.
Competing interests: No competing interests