Inertia, innovation, and cross-lagged effects in negative affect and rumination: daily diary study among people living with HIV

ABSTRACT

Objective

The aim of this study was to examine individual differences in the day-by-day relationship between negative affect (NA) and rumination in terms of their inertia, innovation, and cross-lagged effects among people living with HIV (PLWH).

Methods

The participants were 217 PLWH with confirmed diagnoses of HIV and undergoing antiretroviral treatment. They assessed their NA and rumination for five consecutive days each evening via an online survey.

Results

Results showed that inertia in NA is negatively related to inertia in rumination. Both innovations were unrelated. However, the individuals with relatively higher overall NA were also more reactive to external factors and/or had more variability in their daily lives, to which they respond with NA. Finally, the autoregressive effects were revealed to be important for spillover effects in a direction that is coherent with a given inertia. Thus, the direction of the cascade between daily NA and rumination depends on the area of major regulatory weakness.

Conclusion

The results support the view that intensity, inertia, and innovation are distinct dimensions in spite of the common assumption that higher overall intensity of emotions and coping should be strongly related or even synonymous to their perseveration.

KEYWORDS:

• Emotional inertia
• coping flexibility
• negative affect
• rumination
• HIV/AIDS

The relationship between emotions and psychological well-being has been thoroughly examined in dozens of studies that involved versatile samples and numerous research contexts (e.g., reviews and meta-analysis: Fredrickson & Joiner, 2002; Hülsheger & Schewe, 2011; Lyubomirsky et al., 2005; Watson et al., 1988). Therefore, invoking this subject again requires a convincing justification. One of such rationales deals with the methodological limitations shared by majority of studies from that field, which refer collectively to the static perspective on emotions (see meta-analysis: Houben et al., 2015). Alternatively, the traditional outlook on emotions was based on measuring them as single emotional states that can be experimentally turned on and off, or as stable individual traits. This was a predominant approach despite the fact that emotions have been defined in classic theories of emotions as phenomena characterized by inherent dynamics in time, and not in terms of static entities (Frijda, 1988). It seems that the advancements in theoretical models of emotions preceded the progress in the empirical ways of measuring them, stemming mostly from the limitations in the available statistical methods at that time (Houben et al., 2015). The number of research using an intensive longitudinal design (Bolger & Laurenceau, 2013) has been increasing rapidly, with a stronger focus on the temporal characteristics of the processes revealed by the data gathered in the natural environments of the participants (e.g., Hamaker & Wichers, 2017; Larsen et al., 2009; Scherer, 2009). Those studies showed that emotions should be operationalized as emergent processes since short-term affect dynamics translates into various positive or negative long-term well-being outcomes (Koval et al., 2016; Wichers, 2014). One of the parameters used to describe such dynamics is called emotional inertia. This concept was first introduced by Suls et al. (1998) to capture how a current emotional state may be predicted from an individual’s previous emotional states. More precisely, this term describes an autocorrelation between two consecutive emotional states; being an indicator of emotional flexibility, it also functions as a predictor of psychopathology (Houben & Kuppens, 2020; Kuppens et al., 2010, 2012). Emotional inertia can be operationalized in models characterized by a repetitive measurement of the same set of variables for a given person within relatively short periods of time. Yet there is a dearth of research on its accompanying phenomenon, namely, emotional innovation, which illustrates what is new in an individual’s emotions on a given day. It is the part of emotional variance that cannot be explained by autoregression (Jongerling et al., 2015).

Similar to the relationship between emotions and well-being, extensive and highly heterogeneous studies on coping have been unable to provide a satisfactory answer on the fundamental question of how to measure coping to cover all of its complexity (see for review: Aldao et al., 2010; Richardson et al., 2017; Skinner et al., 2003). It seems that the classical Lazarus and Folkman (1984) theoretical model, which highlighted that coping is not a unitary, easy to observe behavior, but rather a multidimensional phenomenon consisting of dynamic cognitive, emotional, and behavioral efforts displayed by the same person simultaneously, was way ahead of the statistical and methodological possibilities of its time, leaving it open to charges of being empirically unverifiable (Cheng et al., 2014). One of the main impediments in coping studies is the retrospective approach to coping assessments that is devoid of ecological validity and mainly comprises coping dynamics (Kato, 2013). To overcome this obstacle, an increasing number of studies, starting with a classical paper by Stone et al. (1995), have implemented an intensive longitudinal design to underpin fluctuations in coping in an individual’s natural life (Gunthert & Wenze, 2012). This approach enables the capture of a variability of coping, not only between people, but also within people (Bolger et al., 2003). It may even go a step further, that is, it leads to an understanding that intrapersonal variability may also have the status of individual differences, as described by the concepts of inertia and innovation (Hamaker, 2012; Suls et al., 1998). In this perspective, a flexibility of coping can be analyzed in a brand-new conceptualization. Traditionally, it is an active adaptation of coping strategies to changing conditions, thereby enabling coping effectiveness (Cheng & Cheung, 2005). In most advanced studies in this field, it was examined using the goodness of fit hypotheses, where a given coping behavior should match the characteristics of a situation, with controllability as the main element (Finkelstein-Fox & Park, 2019). Although very promising, this method still examines concurrent associations, and not between-person differences, in terms of the ability to change coping accordingly. This can be approximated by the inertia and innovation concepts. Flexibility refers to the dynamics of coping, namely, the changes in this process from one moment to another. If an individual has a high autocorrelation for a coping strategy, it means there is a lack of flexibility in this regard or, differently speaking, high inertia. On the contrary, a coping innovation would be an illustration of what is new in coping on a given day, i.e., this part of its variance, which cannot be explained by inertia (Jongerling et al., 2015). As these are parameters of an individual’s stability of coping behaviors across time, a high inertia in coping is conceptually similar to a coping style, a trait-like preference to use a given set of coping strategies. In this meaning, a coping style could be defined not only as a frequency and intensity of coping, but also as a level of regulatory weakness. In particular, as there are some evidence proving that an effect on a given strategy depends on who uses this strategy: the use that is untypical to the preference may modify an emotional state to a greater extent (Gruszczyńska & Knoll, 2015).

Major advancements in the treatment of HIV infection have altered the social outlook on HIV/AIDS from a definitely terminal condition to a chronic but manageable illness (Carrico, 2019). Nevertheless, people living with HIV (PLWH) still suffer from high levels of psychiatric disorders, among which depression is the most prevalent (Tran et al., 2019). In addition, the aforementioned few studies on the daily functioning of PLWH found that in spite of the same source of distress (i.e., HIV infection) the PLWH’s distress level may dynamically change day-by-day, leading to substantial individual differences in psychological adjustment over time. Some studies attribute these differences to emotion dysregulation, defined as difficulty in the self-regulation of one’s affective states and emotion-driven behaviors (Brandt et al., 2017) stemming from the struggle with internalized HIV-related stigma (Rendina et al., 2018). Emotion dysregulation is present mostly in newly diagnosed PLWH (Bhatia et al., 2011), but is also related to maintaining high levels of negative affect many years after HIV diagnosis (Do et al., 2014). Conversely, it should be underscored that other studies on coping with HIV infection showed that long-infected PLWH may display affective adaptation, especially if the HIV infection is properly controlled through treatment (Moskowitz et al., 2017). Examining inertia and innovation in both negative affect and rumination may shed some new light on these inconclusive results.

Current study

In the current study, we focus on the dynamic interplay over time between emotion and coping, adopting a model based not on concurrent relationships, but on their variability, expressed by inertia and innovation. That is, we examine the association between negative affect (NA) and rumination among PLWH. Although this link has been long and extensively studied mostly with respect to affective disorders (see for review: DeJong et al., 2016; Nolen-Hoeksema et al., 2008), the majority of studies followed a cross-sectional design. Even when a longitudinal framework was used (Michl et al., 2018), it was still without a systematic focus on random effects, describing between-person differences in the above relationship (Hamaker et al., 2018). To date, only Koval et al. (2012) studied emotional inertia and rumination in a non-clinical sample, and found that they are positively, yet independently, related to depression. However, that study focused on only one trait-like assessment of rumination at the beginning. Thus, real-life inertia in this regulatory process was not taken into account.

To formulate the hypotheses, the adopted model requires a more detailed description. When measuring negative affect (NA) and rumination day-by-day, the relationship between them for a given person can be illustrated with:

1. The mean values of NA and rumination that reflect their trait-like characteristics across all of the measurement points. These serve as an equilibrium over which daily affective states, as well as intensity of rumination, fluctuate. Thus, these values refer to a trait-like level, describing individual differences in NA and rumination.

2. The autoregressive effects for NA and rumination that reflect their emotional inertia and coping inertia, respectively. They constitute a carryover effect and indicate how quickly a person restores their equilibrium. The higher the inertia, the longer it takes, which suggests lower flexibility.

3. The cross-lagged effects from previous-day NA to next-day rumination and, analogically, from rumination to NA, which reflects a potentially causal mechanism between these two variables. Similar to traditional cross-lagged models, these spillover effects present how the changes in one domain affect the changes in another domain day-by-day.

4. The residuals that reflect innovation in NA and rumination. These are parts of the variance that are not explained by their respective autocorrelation and spillover effects.

5. A covariance of these innovations.

Based on existing research (Koval et al., 2012), we assumed that inertia in NA is positively related to inertia in rumination among PLWH (Hypothesis 1), but also that innovation in NA is positively related to innovation in rumination (Hypothesis 2). These two patterns of reactions to daily hassles are likely to co-occur, as both are characterized by personal flexibility. Thus, they may share the same underlying personality base (Bolger & Schilling, 1991). Even if a tendency to ruminate describes a self-perceived perseverance in using this coping strategy (Nolen-Hoeksema, 2000), an inertia, i.e., a regulatory weakness in extinguishing it for a longer period of time from the event triggering it, can be regarded as a pathological core. Therefore, a person may tend to ruminate, but can still master ways to reduce its carryover effect (Watkins, 2009). There is also ongoing debate if NA is a cause or an effect of rumination. When NA can be a response to a variety of factors, sometimes adequate and adaptive (Frijda, 1988), coping should be at least partially responsible for its effective regulation (Aldao et al., 2010). Moreover, an individual’s tendency to react with NA and to use rumination should modify the resultant direction of this cascade (Schuurman et al., 2016). Thus, the magnitude of a spillover effect may be related to the general level of NA and rumination, namely, a higher NA is likely to enhance daily spillover from NA to rumination, whereas a higher rumination will likely enhance the other direction, i.e., from daily rumination to NA (Hypothesis 3).

Methods

Participants and procedure

The participants were composed of 217 PLWH (83% men) who were medically diagnosed and undergoing antiretroviral treatment in the outpatient clinic where they were recruited. The inclusion criteria were lack of HIV-related cognitive disorders and absence of current substance abuse.

The study protocol was accepted by the Research Ethics Committee of the University of Economics and Human Sciences in Warsaw, where the first author previously worked. For five consecutive days (Monday–Friday), after providing informed consent, the participants filled out online questionnaires that were sent to them via hyperlinks to their email boxes each evening. They assessed their negative affect state and rumination around a central hassle on a given day. To check if participants were focused on one particular source of difficulty each day, they assigned it to one of five categories: health and symptoms of the illness, relationships with other people, professional work, household chores, and others (to be specified). A single online survey took about three to five minutes to fill out. The items within a part of the survey devoted to an evaluation of a given variable were randomly mixed to avoid habitual answering, and automatic notifications facilitated the participants to answer every question. Daily access to the diary was restricted to a given time, i.e., up to seven hours starting from 6 pm, after which the link became inactive. This access time was established on the basis of a pilot study and takes into account individual differences in the typical daily schedule as the participants were instructed to complete the questionnaires shortly before going to bed. There was no possibility to modify already sent answers. The participants were not remunerated for their participation.

Measures

The negative affect was evaluated using six items (upset, afraid, tired, unhappy, angry, sad) from the PANAS-X by Watson and Clark (1994). The participants assessed how they felt at the end of each day and provided their answers on a five-point scale from 1 = very slightly or not at all to 5 = strongly. The raw values for a given day were added and averaged. The multilevel reliability was assessed with an omega coefficient (Geldhof et al., 2014). The coefficient values were satisfactory at the within- (ω_w = .81) and between-person (ω_b = .97) levels.

Rumination was assessed using two items taken from the Response Styles Questionnaire (Treynor et al., 2003) for rumination (I’ve been thinking about what I’ve been doing to deserve this; I’ve been wondering why I have problems that other people don’t have). They were rephrased to match the daily evaluations. The participants were instructed to provide their answers on a five-point scale from 1 = I haven’t been doing this at all to 5 = I’ve been doing this a lot, keeping in mind their coping with a central hassle. The raw values for a given day were added and averaged, with higher values indicating higher rumination. The reliability of measurement was satisfactory, with lower values for within-person omega coefficient (ω_w = .63; ω_b = .98).

Data analysis

We used the dynamic structural equation modelling (DSEM) described by Hamaker et al. (2018). That is, we tested a two-level bivariate cross-lagged model with random intercepts, random slopes, random residual variance, and a random residual covariance (model 2 in Hamaker et al. (2018) and an extension of example 9.32 in Muthén and Muthén (1998–2018)). It decomposes intensive longitudinal data into within- and between-person parts. In the within-person part, emotional inertia is defined as a random slope expressed by linear regression of NA on a previous-day NA (φ_NN). Analogically, the coping inertia for rumination (φ_RR) is obtained. Next, the random cross-lagged effects from a previous-day NA to rumination (φ_NR) and from a previous-day rumination to NA (φ_RN) are established as a predictive relationship through linear regression. Finally, an innovation in both NA and rumination is modeled as a residual variance. More specifically, the logs are used to ensure their positive values and allow for correlation with other parameters (for details, see: Hamaker et al., 2018). To estimate a covariance between these residuals, a latent factor was created with the factor loadings for NA and rumination fixed at one. The random residual covariance is defined as the log of variance in this factor. The first represents the common part of both innovations (log (Ψ)) and the latter a unique part for each of them (log (π_NA) and log (π_RU)). In the between-person part, these seven fixed within-person parameters together with the within-person means of NA (µ_NA) and rumination(µ_RU) are understood as individual differences with random effects included. All of these parameters are allowed to correlate with each other.

When reporting standardized values, they comprise the average of the standardized values across clusters for each parameter. It means that they are first standardized per person, and then these values are averaged. All of the DSEM analyses were performed using the Mplus version 8.2 (Muthén & Muthén, 1998–2018) with Bayesian estimation, including also dealing with missing data (50,000 draws, two Markov Chain Monte Carlo chains, default priors).

Results

The univariate sample statistics are presented in Table 1. Missing data were noted for 327 cases out of 1085 possible measurement points, which is a 30% attrition rate. These missing data can be regarded as missing at random (Little’s MCAR, χ²= 172.9, df = 152, p = .19).

Table 1. Univariate higher-order moment descriptive statistics.

Variable	Mean	SD	Skewness	Kurtosis	Range
NA	1.70	0.88	1.05	0.24	1–5
RU	2.48	0.91	.10	−0.51	1–5

Note: NA: negative affect; RU: rumination; SD: standard deviation.

Table 2 presents the fixed and random effects in the model. As shown in the table, an inertia for NA is insignificant, which means that in our sample, the overall transition of negative affect from one day to another is close to zero. However, a variance of this parameter can be regarded as substantial, suggesting that individual differences are observed. Thus, for some people, this autoregressive effect can still be significant and, more importantly, it may have different directions. The averaged standardized value for inertia in NA was equal to 0.09 (95% CI = [−0.20, 0.22]). This proved to be small and diversified in terms of direction, especially in comparison to an averaged standardized inertia in rumination, which was equal to 0.24 (95% CI = [0.07, 0.40]) and was positive.

Table 2. Fixed and radom effects obtained in dynamic structural equation modeling of relationship between negative affect and rumination.

Variable	Fixed effects (means)			Random effects (variances)
	Estimate	SD	95% CI	Estimate	SD	95% CI
μNA	1.59	0.06	[1.41, 1.63]	0.14	0.04	[0.09, 0.23]
μRU	2.59	0.09	[2.43, 2.77]	0.37	0.09	[0.22, 0.56]
φNN	0.07	0.08	[−0.09, 0.22]	0.26	0.06	[0.16, 0.41]
φRR	0.25	0.09	[0.07, 0.44]	0.31	0.08	[0.18, 0.49]
φNR	−0.01	0.10	[−0.18, 0.21]	0.32	0.10	[0.18, 0.55]
φRN	0.12	0.08	[−0.04, 0.28]	0.21	0.06	[0.11, 0.35]
log (πNA)	−2.03	0.19	[−2.41, −1.65]	4.37	0.68	[3.20, 5.86]
log (πRU)	−1.10	0.11	[−1.33, −0.88]	0.46	0.18	[0.23, 0.88]
log (ψ)	−5.34	0.34	[−6.02, −4.67]	0.16	0.16	[0.20, 0.63]

Note: μ_NA and μ_RU: random means for daily negative affect and rumination; φ_NN and φ_RR: random autoregressive regression coefficients for daily negative affect and rumination; φ_NR and φ_RN: cross-lagged regression coefficients between daily negative affect and rumination; log(π_NA) and log(π_RU): random log of the variance of the unique parts of the innovations in negative affect and rumination; log(ψ): random log of covariance between the innovations; SD: standard deviation; 95% CI: 95% credible intervals.

Similarly, both cross-lagged effects are insignificant in the sample, but with observed variability across persons. The average standardized effect for a predictive relationship from NA to rumination was −0.04 (95% CI = [−0.23, 0.11]), whereas from rumination to NA, it was 0.06 (95% CI = [−0.05, 0.15]). The within-person averaged proportion of the explained variance was 0.44 (95% CI = [0.40, 0.50]) for NA and 0.31 (95% CI = [0.26, 0.36]) for rumination.

Table 3 presents correlations of random effects at the between-person level, which served as the basis for the hypothesis verification. In our sample, inertia in NA is negatively, not positively, related to inertia in rumination, which is contradictory to the assumption in Hypothesis 1. Thus, a higher carryover in NA corresponds to a lower carryover in rumination.

Table 3. Correlations between standardized values of random effects.

Variable	μNA	μRU	φNN	φRR	φNR	φRN	log (πNA)	log (πRU)
μNA	1
μRU	−.33	1
φNN	−.04	−.20	1
φRR	.19	.32	−.77	1
φNR	−.13	−.39	.87	−.86	1
φRN	−.05	−.02	−.80	.67	−.63	1
log (πNA)	.92	−.37	.12	−.01	.08	−.19	1
log (πRU)	.23	.28	.07	−.10	.07	−.21	.30	1
log (ψ)	.64	−.23	.10	−.01	.09	−.11	.67	.24

Note: μ_NA and μ_RU: random means for daily negative affect and rumination; φ_NN and φ_RR: random autoregressive regression coefficients for daily negative affect and rumination; φ_NR and φ_RN: cross-lagged regression coefficients between daily negative affect and rumination; log(π_NA) and log(π_RU): random log of the variance of the unique parts of the innovations in negative affect and rumination; log(ψ): random log of covariance between the innovations.

Only for bolded values 95% credible interval does not contain zero.

Similarly, as innovations in NA and rumination were independent, there was no confirmation for Hypothesis 2 in the data. Further analyses of other correlations suggest that individuals who are relatively higher in overall NA are also more reactive to external factors or have more variability in their daily lives, to which they respond with NA, as both explanations are valid in this case (Hamaker et al., 2018). On the contrary, being higher in overall rumination is related to lower innovation in NA, but only in terms of its unique part.

When it comes to Hypothesis 3, overall NA was unrelated to both cross-lagged effects. The same result was obtained for overall rumination. Instead, the autoregressive effects were revealed to be important for the spillover effects. That is, a higher carryover in NA was related to a higher spillover from NA to rumination and lower spillover from rumination to NA. For higher carryover in rumination, the direction was the opposite: higher spillover from rumination to NA and lower from NA to rumination were noted.

Discussion

The obtained findings were quite unexpectedly inconsistent with our hypotheses. More specifically, the inertias in NA and rumination were negatively (not positively) related: a higher carryover in NA was associated with a lower carryover in rumination. Moreover, a higher general level in NA and rumination was unrelated to the spillover effects. Finally, the innovations in NA and rumination were independent of each other. These results are not only counter-intuitive, but mostly in contrast to a large body of literature pointing to the well-recognized vicious circle between negative affect and rumination (e.g., Lyubomirsky & Nolen-Hoeksema, 1995; Nolen-Hoeksema, 2000, 2008). The overall picture seems likely to be more complex than a predisposition of individuals with high NA to ruminate, which, in turn, further stimulates NA (Suls & Martin, 2005). On a daily basis, these processes appear to be interrelated in an advanced manner, as higher inertia enhances the cross-lagged effect in a direction consistent with this inertia and suppresses the opposite one. This holds true for both inertia in NA and inertia in rumination, and sheds light on how carryover effect may be related to spillover effect. Thus, the direction of the cascade between daily NA and rumination depends on the area of major regulatory weakness.

These findings support the notion that intensity of emotion and, analogically, intensity of emotion-focused coping, is not necessarily crucial for well-being, but rather, the previously mentioned lack of flexibility in these processes (Cheng & Cheung, 2005). Again, it is vital to go back to basics, i.e., to Lazarus and Folkman’s (1984) stress and coping theory, which posits that there is no single coping strategy that is definitively adaptive or maladaptive across all stressful situations. Several studies have shown that coping flexibility may be related to lower levels of depression and anxiety (Lougheed & Hollenstein, 2012) or better adaption to traumatic events (Galatzer-Levy et al., 2012), and might even facilitate post-traumatic growth (Cohen & Katz, 2005). However, the central problem with this concept is not only the scarcity of prospective studies, which precludes analyzing its temporal dynamics (Bonanno & Burton, 2013), but operationalizing this construct only at a trait-like level with a single measurement, which ignores its idiosyncratic character, varying within the individual, situation, and time (Lazarus & Folkman, 1984). To our knowledge, this study is the first example to utilize inertia and innovation in the context of coping.

Additionally, the findings show that the trait-like levels of NA and rumination were unrelated to inertia effects and were significant for innovation effect only in NA. In other words, a higher NA and a higher rumination do not necessarily translate into higher carryover effects, but may modify “fresh” reactions in NA. This supports the view that intensity and inertia are distinct dimensions in spite of the common assumption that higher overall intensity of emotions and coping should be strongly related or even synonymous to their perseveration. In fact, intense but limited-time reactions may have a different impact on well-being than less intense but long-lasting reactions. Similarly, an individual’s sensitivity, understood as the strength of a new response, i.e., not explained by autoregressive and spillover effects, cannot be straightforwardly deduced from a general trait-like level of a given variable. In this light, although speculative, neuroticism, for instance, could be described as a specific combination of three dimensions, namely, intensity, inertia, and innovation, assuming co-existence for the same person across many situations high level of NA, high carryover of NA, and high innovation of NA. Analogously, a pathological ruminative style should consist of a combination of high intensity of rumination, low flexibility, and high sensitivity to react with rumination. Taken together, it calls for more in-depth and multidimensional characteristics of both phenomena, namely, affect and coping, especially in the face of ambiguous results concerning their relationship (Aldao et al., 2010). Perhaps, at least to some degree, an explanation lies in a parametrization of the processes, with temporal descriptors as crucial factors in covering the dynamics of real life (Bos et al., 2019).

Finally, it is truly challenging to discuss the obtained results in the context of the uniqueness of our sample consisting of PLWH. It is not only due to an absence of studies on emotional inertia among PLWH specifically, but also due to very scarce research on this phenomenon in the clinical context (Dejonckheere et al., 2019). Struggling with somatic illness poses chronic psychological distress, but several studies have pointed out the affective adaptation to such conditions according to the hedonic treadmill model (Lyubomirsky, 2010), which describes a “stability despite loss” of well-being in reaction to disease or other critical life event (Schilling & Wahl, 2006). In the case of PLWH, several authors have observed that despite HIV-related distress, PLWH’s positive and negative affect are remarkably stable even many years after their HIV diagnosis; they are also not associated with the HIV-related clinical variables, at least in the samples with good control of infection progression through antiretroviral treatment (e.g., Carrico & Moskowitz, 2014; Moskowitz et al., 2017; Rzeszutek & Gruszczyńska, 2018). Thus, individual differences in this regard should be directly related to the mechanisms based on emotional, cognitive, and behavioral flexibility.

In this perspective, a relationship between emotional inertia and coping inertia may be an underlying factor of depression in these patients (Moskowitz et al., 2009, 2017; Rendina et al., 2018), a possibility that needs further research. The finding that for PLWH with higher inertia in NA, the spillover is from NA to rumination, whereas for those with higher inertia in rumination, that direction is the opposite, is not only theoretically relevant but may be clinically meaningful for creating better targeted interventions in this patient group (Moskowitz et al., 2017; Wilson et al., 2016).

Strengths and limitations

Our study has several strengths, including a theory-driven, dynamic, multilevel modeling of daily functioning in a clinical sample of PLWH. The major novelty of the study lies in the use of the concept of inertia and innovation to examine coping. Furthermore, surprisingly few studies so far have concerned daily rumination, especially outside the context of depressive symptoms. However, some limitations should also be underscored. First, there was a low number of measurement points, which, apart from sample uniqueness, can shape the obtained findings. Therefore, it is reasonable to look at it as a preliminary study. Second, we were not able to avoid sample selection bias that is typical for diary studies as it can be assumed that only highly functional PLWH were accessible during the recruitment process. Finally, all results should be treated with caution as they omit other variables influencing daily functioning. In addition, the time lag can be regarded as perhaps too long for a valid reflection of dynamics. Thus, the results should be confirmed in ecological momentary assessment with a few measurements on each day, varying randomly across and within individuals (Hamaker & Wichers, 2017).

Conclusion

The results of our study undermine the traditional outlook on the existence of a vicious circle between NA and rumination, examined here for the first time in the clinical context of living with HIV/AIDS. These constructs are probably interrelated in a more complex manner, including temporal parameters of affect and coping as a process, just like inertia, innovation, and spillover effects, which should be taken into account in both theory and practice. In a broader perspective, it is also important to note that people differ in their ability to identify and report small fluctuations in their emotional, cognitive, and behavioral responses (Houben & Kuppens, 2020). These individual differences may influence results either directly or indirectly through interaction with a research method and design. Hence, this problem occurs to some extent in any self-descriptive study, and we have no reason to assume that our study is particularly affected by it. However, the question of whether such self-awareness modifies daily or momentary reports and – through them – inertia and innovation, is worth further investigation.

When planning interventions probably greater focus would be needed not only on intensity and adequacy of, cognitive, emotional and behavioral responses (Campbell-Sills & Barlow, 2007), but also directly on their immediate flexibility, making it a part of systematic diagnosis and targeted actions. Recently some authors showed the promising role of mindfulness in the aforementioned process, i.e., promoting adaptive patterns of affective changes by stimulating adaptive coping in daily life (Keng & Tong, 2016). This approach may be of particular relevance to PLWH, as daily struggle with chronic and incurable disease in some of them leads to dysregulation and maladaptive response patterns, also as a result of HIV/AIDS stigma (Rendina et al., 2018).

Research involving human participants

The study protocol was accepted by the institutional ethics committee (decision made 2016/06/06). Written informed consent was obtained from all participants before participation in the study.

Acknowledgements

The authors wish to thank Ewa-Firląg Burkacka, PhD, from the Warsaw’s Hospital for Infectious Disease for help in participants recruitment.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

All the data are available upon the request from the corresponding author.

Additional information

Funding

This work was supported by National Science Center, Poland (Narodowe Centrum Nauki): [grant number: 2016/23/D/HS6/02943].